Old Calls

The objective of this thesis is to advance the mathematical theory of Bayesian Neural Networks for both Shallow and Deep Networks. These kind of networks are very promising since their flexibility in modeling the phenomenon under study as well as to provide accuracy measures of the estimated quantities a problem not well understood in classic network theory. Computational aspects would another aspect cover by this thesis since standard methods for network training will not be suitable for real applications. The other objective of this thesis is the applications of the new methodology to important problems in Health Sciences for e.g. Omics or  Bioimaging data. The study will involve expertise in several topics including Bayesian non parametrics methods, information geometry, Monte Carlo methods as well as standard methods in Deep Learning. The developed methods will be tested on different life sciences datasets (separatly and integrated) such as EHR, Omics and bioimaging (e.g. CT, PET, digital pathology), both publicly available and originally produced by DSH partner labs, on different tasks.

The objective of this thesis is to advance mathematical models and analytical tools for the investigation of spatially embedded networks and their influence on urban systems. The study will utilize the expertise of several fields, including graph theory, topology, geometry, and spatial statistics, to analyze the structure, function, and dynamics of spatial networks.

Through our research, we aim to contribute to the comprehension of urban systems by developing appropriate mathematical and statistical tools with the goal of providing insightful information to policy decisions concerning urban planning and transportation. Our study will focus on (i) theoretical properties of spatially embedded networks (ii) their design for modeling urban systems and (iii) the statistical tools for their analysis.

To achieve these objectives, we will adopt a multidisciplinary approach that will advance our understanding of spatially embedded networks as a mathematical model and their use in the analysis of urban systems. Our focus will be on the development of mathematical models that effectively capture the complexity of urban spatial networks such as transportation and utility networks. The analysis will try to identify patterns, predict outcomes, and classifying spatial networks based on their properties.

Computational issues are an aspect that will be covered by the project as well. By combining traditional statistical methods with machine learning algorithms, we will be able to process large networks to extract insights that would be otherwise challenging to obtain. Moreover, machine learning techniques will be utilized to optimize network design and allocate resources, resulting in more sustainable and efficient urban systems.

Candidates will learn and apply mathematical modeling techniques for the study of infectious disease dynamics and control, with a special focus to their deployment within important public health issues. Compartmental modeling, individual-based modeling and Bayesian approaches will be among the main tools adopted to this aim. Infections due to respiratory or mosquito-borne pathogens and antimicrobial resistant microorganisms will be among the main possible topics of the research. Potential applications are the quantitative estimation of critical parameters and their population heterogeneity, the understanding of mechanisms of transmission and disease, the assessment of effectiveness of public health interventions and the development of epidemiological scenarios, including those relevant for pandemic preparedness and response. The PhD will take place at the Center for Health Emergencies (CHE) of the Fondazione Bruno Kessler under activities of the PNRR-funded INF-ACT project, in which the CHE leads the development of mathematical modeling activities within a partnership involving most of the major Italian research and public health institutions in the field of infectious diseases.

Conversational agents are experiencing a surge in interest given the continuous release of new models and the ever evolving scenario of NLG. Still, the actual focus is mainly on model size, training data size and prompt engineering. The interaction of these elements with related aspects, such as decoding strategies, knowledge guided generation, data quality, knowledge distillation -just to mention a few- can help in improving the models, especially for better factuality, reducing hallucination and increasing coherence among dialogue turns. The goal of this PhD Thesis is to overcome the shortcomings of present large language models by incorporating novel strategies for better generation.

This research aims to develop an innovative framework that combines 3D metrology, Internet of Things (IoT) and blockchain technologies for secure and trustworthy data management in real-time monitoring of civil infrastructures. The research will investigate the integration of IoT-enabled sensors and advanced 3D surveying techniques (SLAM, LiDAR, etc.) for continuous monitoring of structural health whereas blockchain technology will support secure and decentralized data storage and sharing.

The interdisciplinary Phd will contribute to advancing IoT, 3D metrology and blockchain technologies and their integration, providing a secure and reliable framework for real-time monitoring of civil infrastructure, such as bridges, buildings, dams or monument. The research findings will therefore be applicable to various industries, including transportation, construction, mining, etc. and could have significant implications for public safety and economic development.

Planning and scheduling are techniques to automate and/or optimize decision-making. There is a breadth of applications that can benefit from the application of this kind of technique including (but not limited to) robotics, flexible manufacturing, logistics and people management.

The aim of this PhD scholarship is to investigate and reinforce the applicability of this kind of technique considering the whole spectrum of domains that recently emerged from the AIPlan4EU (aiplan4eu-project.eu) project. The candidate will research innovative approaches and algorithms to improve the performance, usability and/or relevance of planning and scheduling techniques deployed in diverse scenarios, having the unique possibility to work and experiment with real-world scenarios of planning already deployed by the project.

Industrial systems are reaching an unprecedented degree of complexity. The process of designing a complex system is expensive, time consuming and error-prone. Moreover, the design process has to guarantee not only the functional correctness of the implemented system, but also its dependability and resilience with respect to run-time faults. Hence, the design process must characterize the likelihood of faults, mitigate possible failures, and assess the effectiveness of the adopted mitigation measures.

Formal methods have been increasingly used over the last decades to deal with the shortcomings of designing a complex system. Formal methods are based on the adoption of a formal, mathematical model of the system, shared between all actors involved in the system design, and on a tool-supported methodology to aid all the steps of the design, from the definition of the architecture down to the final implementation in HW and SW. Formal methods include technologies such as model checking, an automatic technique to symbolically and exhaustively analyze all possible executions of the system in the formal model, in order to detect design flaws as early as possible. Model checking techniques have been recently extended to assess the safety and dependability characteristics of the design, and for system certification.

The objective of this study is to advance the state-of-the-art in system design using formal methods. This includes adapting and extending the system design methodology, investigating improved versions of state-of-the-art routines for verification and safety assessment of complex systems, and developing novel extensions to address open problems. Examples of such extensions include novel techniques for contract-based design and contract-based safety assessment, advanced techniques for formal verification based on compositional reasoning, the analysis of the timing aspects of fault propagation, the characterization of transient and sporadic faults, the analysis of the effectiveness of fault mitigation measures in presence of complex fault patterns, and the modeling of analysis of systems with continuous and hybrid dynamics.

This study will exploit the challenges and benchmarks defined in various industrial projects carried out at FBK.

Techniques based on formal methods for the verification and validation of embedded and safety-critical software systems are becoming increasingly important, due to the growing complexity and importance of such systems in every aspect of modern society. Despite the major progress seen in the last twenty years, however, the application of formal methods in embedded software remains a challenge in practice, due to factors such as the interplay between computation and physical aspects and the increasing complexity of the software and its configurations.

This project will investigate novel techniques for the application of formal methods to the design, verification, and validation of embedded software, with particular emphasis on safety-critical application domains such as railways, automotive, avionics, and aerospace. The techniques considered will include a combination of automated and interactive theorem proving, satisfiability modulo theories, model checking, abstract interpretation, and deductive verification. Examples of the problems tackled during the project include the formal verification of functional requirements expressed in temporal logics, automated test-case generation, efficient handling of parametric/multi-configuration software systems and product lines, and the verification of software operating in a physical environment, subject to real-time constraints. Importantly, in addition to researching novel theoretical results, a significant part of the project activities will be devoted to the implementation of the techniques in state-of-the-art verification tools developed at FBK and their application to real-world problems in collaboration with our industrial partners.

Planning – devising a strategy to achieve a desired objective – is one of the basic forms of intelligence, with applications in autonomous robotics, logistics, flexible production, and many other fields. Historically, planning research has followed a general-purpose framework: a generic engine searches for the strategy by reasoning on the problem statement. Despite substantial progress in recent years, domain-independent planning still suffers from scalability issues and fails to deal with real-word problems. The alternative is to devise ad-hoc, domain-specific solutions that, although efficient, are costly to develop, rigid to maintain, and often inapplicable in non-nominal situations.

The PhD student will study the foundations of an innovative approach to Planning that will be domain-independent and efficient at the same time. The idea is to adopt a framework based on Reinforcement Learning, where a domain-independent planner is specialized with respect to the domain at hand. This research project will advance the state of the art in planning beyond the “efficiency vs flexibility” dilemma and provide effective techniques to be validated on real-world use-cases.

In modern and heterogeneous learning environments, the one-size-fits-all approach is proven to be fundamentally flawed. Individualization through adaptivity is crucial to nurture individual potential, needs and motivational factors.

The goal of this PhD thesis is to investigate the potential of combining gamification mechanics and adaptive personalized learning, analyzing the impact in terms of students’ achievements, participation and motivation. In particular, the PhD candidate will investigate AI-based theories and techniques for the development and validation of an  open, content-agnostic, and extensible platform for personalized playful learning. The platform will be validated in different formal and informal educational contexts.

The ideal candidate has a background in Computer Science or Cognitive Science. Game design, educational and cognitive psychology, motivation theories, knowledge on designing and conducting experimental studies, experience with quantitative and qualitative data analysis techniques are a plus for the application and should be acquired during the Phd training.

This research aims to investigate the design of secure and interoperable data spaces and data governance frameworks for the management and sharing of agricultural data in the context of Agriculture 4.0. The study will explore the technical and organizational challenges of establishing and governing data spaces, which are virtual environments for managing, sharing, and analyzing data. The research will examine the potential benefits of data spaces in agriculture, such as improving crop yields, reducing environmental impact, and enhancing the overall efficiency of agricultural operations. The study will explore how data governance frameworks can be designed to ensure data privacy, security, and accountability in the agriculture industry. The research will investigate the technical and economic factors that influence the adoption and implementation of these frameworks, including issues related to data quality, interoperability, and standardization. The study will also explore how emerging technologies, such as blockchain, edge computing, and machine learning, can enhance the security and governance of data spaces in agriculture. The research will identify strategies for optimizing these technologies and frameworks to promote innovation, collaboration, and sustainability in Agriculture 4.0.

This research aims to investigate the integration of robotics and IoT technologies with deep learning-based IoT Data analysis techniques in the context of digital agriculture. The study will focus on addressing the challenges and opportunities that arise from the combination of these technologies, with a particular emphasis on the use of deep learning algorithms for the mass collection and analysis of agricultural data provided by IoT-distributed sensors. The research will explore how deep learning can improve the efficiency, sustainability, and productivity of agricultural practices by enabling automatic and semi-automatic training of algorithms for classification/analysis. The study will also examine the technical and economic factors that may influence the adoption and implementation of these technologies in the agricultural sector to identify strategies to optimize their use and impact.

Hybrid systems are formal models combining  discrete and continuous-time dynamic behaviors. They can be found in various applications such as robotics, control systems, cyber-physical systems, and transportation systems. Formal methods for hybrid systems provide a powerful set of techniques for designing, analyzing, and verifying the behavior of complex systems that exhibit both continuous and discrete behaviors. These techniques can be used to ensure the correctness and safety of the system and to detect design flaws and bugs early in the development cycle.

This project will investigate new formal methods to prove properties of hybrid systems integrating model checking, automated theorem, and numerical analysis for control theory. Different aspects of hybrid systems will be considered including temporal properties, diagnosability and epistemic properties, reliability and robustness to faults. Compositional reasoning and proof synthesis will be also considered. The new methods will be implemented and evaluated on industrial benchmarks derived by industrial collaboration of FBK in various application domains such as space, avionics, autonomotive, railyways, and energy.

Social learning and learning of cause-and-effect relationships are key components of human intelligence. The goal of the current PhD thesis is to model and evaluate notions of social learning, social influence, and counterfactual and causal learning in order to improve the performance of a group of AI agents and their ability to produce explainable decisions in scenarios where they have to interact with humans. Additionally, the thesis may also investigate the properties of empirical social networks (for example, sparsity) to organize the topology of communications, interactions, and cooperation that occur between a multiplicity of AI agents.

The ideal candidate will have research interests on multi-agent deep reinforcement learning, complex networks, causal learning. The candidate will have the possibility of working within the ELLIS network and in collaboration with top international universities and research centers. Contacts: [email protected]; [email protected]

SPAD arrays in CMOS technology are single photon detectors providing < 100 ps time resolution. They are used in a wide spectrum of time-resolved imaging systems, such as medical and biomedical imaging (e.g., fluorescence lifetime imaging microscopy, Raman spectroscopy, and diffuse optical tomography), depth sensing for industrial, automotive, space and consumer applications, and quantum imaging (super-resolution microscopy, ghost imaging). These systems rely on the high temporal resolution of SPAD detectors combined with time-to-digital converters to generate precise timestamps of individual photons, which are then processed to extract the required information (lifetime, time-of-flight, etc). In case of large arrays of SPADs, the readout channel becomes the bottleneck of the system, demanding the integration of on chip of power- and area-efficient custom digital signal processors. The objective of this project is to develop hardware friendly algorithms for the specific challenges identified above, test them using existing sensors and then designing custom processors to be integrated on-chip. The student will interact with experts in the fields of single-photon image sensors, and analog/mixed signal integrated circuit design, gaining a unique combination of background knowledge. The expected outcome is the realization of state-of-the-art image sensors and their validation in a real use-case scenario.

The widespread use of social media has transformed the way people interact and communicate with each other. The interactions that occur on social media create a network that facilitates the transmission of information across large groups of people. This network can be used to understand how information flows through society and how it shapes people's beliefs and opinions.

One way to understand the social network that is created through online interactions is to analyze the textual information contained within user profiles and messages. This information can be used to characterize users, identify emerging topics, and extract information about cultural characteristics. By studying the patterns of interactions between users, it is possible to identify their stance on specific topics, communication dynamics and communities sharing the same opinions.

However, understanding the social network alone is not enough to gain a complete understanding of how information is transmitted and how it shapes society. It is also important to consider the physical location of these groups and how it affects their interactions and the diffusion of information across the social network. This is because physical space can play a significant role in shaping social dynamics and the diffusion of information. For example, social movements may arise in response to local events, and these movements may have a different impact in different geographic locations.

The interplay between the structure of the social network and the semantic information that diffuses across it is a complex phenomenon that requires a multidisciplinary approach, and that can be further enriched with geographical information. This PhD thesis will employ various research methods, including data mining, network analysis, Natural Language Processing (NLP), and geographic information systems (GIS), to explore this interplay. By examining the relationship between the social network and semantic information, the research aims to provide insights into how social media shapes society and how it can be used to better understand the dynamics of cultural groups, minorities, and social movements of opinion. Ultimately, the findings of this research may have important implications for fields such as sociology, anthropology, and communication studies, as well as for policymakers and social media companies.

In recent years, there has been a significant increase in demand for secure and efficient systems to process data. In-network computing has emerged as a promising solution for offloading computation tasks, reducing latency, and relieving the workload of connected computing nodes. This technology uses smart network interface cards (smartNICs) to perform computation tasks on the network. However, limited adoption of this technology is due to the maturity of the software stack and related programming models, particularly for security applications.

This PhD scholarship aims to investigate many aspects of enabling in-network computing as a newer paradigm for solving security challenges, including cryptography. After conducting a literature review to identify relevant research and industry efforts in this area, focusing on existing systems such as NVIDIA Bluefield and programming models like DOCA or sPIN, the PhD candidate will identify challenges associated with adopting an in-network computing model for security and propose novel technical solutions. Additionally, the PhD candidate will conduct experimental evaluations to measure the performance and security of the proposed solutions (e.g., cryptographic algorithms implemented on smartNICs) and compare these results with traditional approaches. The findings of this research can guide future research and development in this area and can be applicable to industries that require secure and efficient data processing.

The advent of foundation models has introduced unprecedented opportunities in all areas of natural language processing. Automatic translation (be it speech or text translation) is no exception, with a wide variety of language directions, domains and application scenarios whose coverage is no longer a mere utopia. Although conditions today are more favorable than in the past, open challenges still exist in terms of fully exploiting the power of the available models,  increasing their flexibility to integrate diverse input types, or constraining the output to meet specific application requirements. Open questions include: how to feed non-symbolic models with symbolic information describing the context of a translation request? How to supply meta-information about target users? How to integrate model capabilities with external information from structured knowledge bases? How to condition the output to specific target applications? This PhD aims to explore state-of-the-art solutions to tackle these challenges, with a special focus on the integration of multimodal information (e.g. contextual information supplied as visual cues), user-specific constraints (e.g. for gender/formality control), and application-specific constraints (e.g. structural requirements as in the case of video subtitling).

Machine learning and deep neural networks have proven to be highly effective in processing multimodal data, such as audio, video, and environmental data, on powerful computing systems. However, a major challenge in artificial intelligence is how to extend these capabilities to resource-constrained devices, such as end nodes, in an IoT system. Fortunately, recent advancements in TinyML approaches are opening up new opportunities to bring AI to the far edge of the edge-to-cloud continuum. Exciting research scenarios are emerging that span from tiny deep learning solutions for inference on resource-constrained platforms, which rely on distillation, quantization, or neural architecture search, up to combining software techniques with innovative hardware that support TinyML. Moreover, the complexity grows when we consider moving learning to the edge to take advantage of the opportunities presented by connected, distributed devices. To address these challenges, this research aims to (i) develop novel hardware and software approaches for optimizing AI on energy-efficient embedded devices, with a particular focus on audio processing and computer vision, but not limited to these areas; (ii) explore the potential of distributing and fusing intelligence from heterogeneous nodes in an IoT; and (iii) demonstrate the benefits of these approaches in real-world application scenarios, such as those found in smart cities. The candidate's profile and interests will be considered when structuring this interdisciplinary research at the intersection of Artificial Intelligence, Embedded Systems, Distributed Computing, and Low power hardware, contributing to the development of innovative solutions for real-world challenges. The candidate will have the opportunity to work on cutting-edge technology, gain experience in interdisciplinary collaboration, and make significant contributions to the field of tiny machine learning.

In spite of the recent progress in speech technologies, processing and understanding conversational spontaneous speech is still an open issue, in particular in presence of challenging acoustic conditions as those posed by dinner party scenarios. Although enormous progresses have been made recently in a variety of speech processing tasks (such as speech enhancement, speech separation, speech recognition, spoken language understanding), a unified established solution is still far from being available. In particular, one of the limitations of the current approaches is their computational complexity that makes an actual deployment in low-end or IoT devices not feasible in practice.

The candidate will advance the current state-of-the-art in speech processing (in particular for separation and enhancement) towards developing a unified solution, possibly based on self-supervised or unsupervised approaches, for automatic speech recongition in dinner party scenarions (as those considered in the CHiME challenges).

In 2023 there will be the launch of the European Space Agency (ESA) mission JUpiter ICy moons Explorer (JUICE) to the Jovian system, and the development of the ESA EnVision mission to Venus will continue. Both missions carry on board a radar sounder instrument for subsurface sensing of planetary bodies. In the context of these two projects we are looking for candidates willing to design methodologies for sub-surface radar image processing and analysis. The outcome of this activity will contribute in improving the understanding of subsurface structures, and their correlation to planetary body history and climate.

The candidate will be requested to design and develop novel methodologies based on machine learning, deep learning, pattern recognition and artificial intelligence for information extraction, classification, target detection, noise reduction and change detection in radar and radar sounder images.

Besides the requirements established by the rules of the ICT school, preferential characteristics for candidates for this scholarship are:

• master degree in Electrical Engineering, Communication Engineering, Computer Science, Mathematics or equivalents;

• knowledge in pattern recognition, deep learning, image/signal processing, statistic/remote sensing/radar.

Nowadays a huge amount of remote sensing data is available being acquired with a high temporal, spatial and spectral resolution. Those data are coming from several missions, among the others: ESA Copernicus (Sentinels), ASI PRISMA and COSMO-SkyMed, and future IRID constellation. The management and use of such data requires the design of novel solutions being able to handle long time series of dense but irregular data worldwide in the context of high-power computing systems looking both back and forth in time.

Candidates will be requested to develop novel methodologies based on machine learning, deep learning, pattern recognition and artificial intelligence for information extraction, classification, target detection and change detection in long and dense time series of remote sensing images.

Besides the requirements established by the rules of the ICT school, preferential characteristics for candidates for this scholarship are:

• master degree in Electrical Engineering, Communication Engineering, Computer Science, Mathematics or equivalents;

• knowledge in pattern recognition, deep learning, image/signal processing, statistic/remote sensing, passive/active sensors.

This PhD position calls for research into "foundation models," a type of deep learning neural network. These models, trained using vast amounts of data, serve as starting points for various tasks including, but not limited to, classification, regression, segmentation, and detection. However, the application of these models to the understanding of 3D scenes presents a challenge due to the unique characteristics of the training data. The proposed research aims to bridge this gap. The PhD candidate will primarily be tasked with exploring and constructing innovative vision algorithms rooted in deep learning and based on foundation models. The goal is to develop algorithms that can be effectively and flexibly utilised across diverse data types and domains.

The Internet of Things (IoT) is enabling and multiplying the point of collection of multimodal data such as audio, video and environmental data, typically processed in the cloud where potentially infinite computational power is available. However, this comes at the cost of bandwidth, energy and privacy. Recent research in the so-called tinyML domain is tackling the challenge of bringing artificial intelligence to the end-devices, thus limiting the need to stream data to the cloud and implementing the distributed intelligence paradigm in the cloud-edge continuum. Thus, novel approaches to enable AI, typically computationally demanding, on resource-limited devices are needed.  Exciting research scenarios are emerging to enable inference at the edge, ranging from distillation, quantization, or neural architecture search strategies to the fusion of software techniques with innovative hardware supporting tinyML. The complexity grows when we consider shifting not only inference but also learning to the edge to harness the opportunities offered by connected, distributed devices. The research proposed will focus on one or more of the following goals: (i) Develop novel hardware and software approaches to optimize AI on energy-efficient embedded devices, with a particular emphasis on audio processing and computer vision, while also considering other domains. (ii) Explore the potential of distributing and integrating intelligence from heterogeneous nodes in an IoT environment. (iii) Demonstrate the benefits of these approaches in real-world application scenarios, such as those encountered in smart cities.

This interdisciplinary research at the intersection of Artificial Intelligence, Embedded Systems, Distributed Computing, and Low-power Hardware will take into account the candidate's profile and interests, contributing to the development of innovative solutions for real-world challenges. The candidate will have the opportunity to work with cutting-edge technology, gain valuable experience in interdisciplinary collaboration, and make significant contributions to the field of tiny machine learning.

A better understanding of Earth climate and evolution relies on a better understanding or other planets. In 2023 European Space Agency (ESA) launched the JUpiter ICy moons Explorer (JUICE) mission to the Jovian system, and the development of the ESA EnVision mission to Venus is under development. Both missions carry on board a radar sounder instrument for subsurface sensing of planetary bodies. In the context of these two projects we are looking for candidates willing to design methodologies for sub-surface radar image processing and analysis. The outcome of this activity will contribute in improving the understanding of planetary subsurface structures, and their correlation to history and climate, as well as of Earth.

The candidate will be requested to design and develop novel methodologies based on artificial intelligence, deep learning and machine learning for information extraction, segmentation, classification, target detection, noise reduction and change detection in radar and radar sounder images.

Besides the requirements established by the rules of the ICT school, preferential characteristics for candidates for this scholarship are:

• master degree in Electrical Engineering, Communication Engineering, Computer Science, Mathematics or equivalents;

• knowledge in pattern recognition, deep learning, image/signal processing, statistic/remote sensing/radar.

In recent years, there has been a significant increase in demand for secure and efficient systems to process data. In-network computing has emerged as a promising solution for offloading computation tasks, reducing latency, and relieving the workload of connected computing nodes. This technology uses smart network interface cards (smartNICs) to perform computation tasks on the network. However, limited adoption of this technology is due to the maturity of the software stack and related programming models, particularly for security applications.

This PhD scholarship aims to investigate many aspects of enabling in-network computing as a newer paradigm for solving security challenges, including cryptography. After conducting a literature review to identify relevant research and industry efforts in this area, focusing on existing systems such as NVIDIA Bluefield and programming models like DOCA or sPIN, the PhD candidate will identify challenges associated with adopting an in-network computing model for security and propose novel technical solutions. Additionally, the PhD candidate will conduct experimental evaluations to measure the performance and security of the proposed solutions (e.g., cryptographic algorithms implemented on smartNICs) and compare these results with traditional approaches. The findings of this research can guide future research and development in this area and can be applicable to industries that require secure and efficient data processing.

Applications are invited for a Ph.D. opportunity focused on secure data spaces for trustworthy data sharing in digital agriculture. The research aims to develop robust platforms for data sharing, privacy, and integrity in the agricultural landscape. Applicants should have a background in computer science or data management, with familiarity in agricultural systems and/or practices, data analytics, machine/deep learning, and algorithm design, validation, and benchmarking.

Techniques based on formal methods for the verification and validation of embedded and safety-critical software systems are becoming increasingly important, due to the growing complexity and importance of such systems in every aspect of modern society. Despite the major progress seen in the last twenty years, however, the application of formal methods in embedded software remains a challenge in practice, due to factors such as the interplay between computation and physical aspects and the increasing complexity of the software and its configurations.

This project will investigate novel techniques for the application of formal methods to the design, verification, and validation of embedded software, with particular emphasis on safety-critical application domains such as railways, automotive, avionics, and aerospace. The techniques considered will include a combination of automated and interactive theorem proving, satisfiability modulo theories, model checking, abstract interpretation, and deductive verification. Examples of the problems tackled during the project include the formal verification of functional requirements expressed in temporal logics, automated test-case generation, efficient handling of parametric/multi-configuration software systems and product lines, and the verification of software operating in a physical environment, subject to real-time constraints. Importantly, in addition to researching novel theoretical results, a significant part of the project activities will be devoted to the implementation of the techniques in state-of-the-art verification tools developed at FBK and their application to real-world problems in collaboration with our industrial partners.

In modern and heterogeneous learning environments, the one-size-fits-all approach is proven to be fundamentally flawed. Individualization through adaptivity is crucial to nurture individual potential, needs and motivational factors. The goal of this PhD thesis is to investigate the potential of combining gamification mechanics and adaptive personalized learning, analyzing the impact in terms of students’ achievements, participation and motivation. In particular, the PhD candidate will investigate AI-based theories and techniques for the development and validation of an  open, content-agnostic, and extensible platform for personalized playful learning. The platform will be validated in different formal and informal educational contexts. The ideal candidate has a background in Computer Science or Cognitive Science. Game design, educational and cognitive psychology, motivation theories, knowledge on designing and conducting experimental studies, experience with quantitative and qualitative data analysis techniques are a plus for the application and should be acquired during the Phd training.

Social learning strategies are a key component of human intelligence and of our ability of learning from and collaborating with other humans in their environment. Inspired by this, some initial research efforts are enabling embodied AI agents with social and cooperative skills, thus permitting them to coordinate and collaborate with and to learn from other agents and humans. The goal of this PhD thesis is to devise innovative cooperative learning algorithms to support navigation of embodied AI agents in dynamic complex environments populated by other agents and humans, and to learn how to interact and to communicate with other heterogenous agents to learn collaboratively world models. The ideal candidate would have research interest in cooperative and embodied AI, multi-agent deep reinforcement learning, neuro-symbolic approaches, etc. The candidate will have the possibility of working within the ELLIS network and in collaboration with top international and national universities.

Planning and scheduling are techniques to automate and/or optimize decision-making. There is a breadth of applications that can benefit from the application of this kind of technique including (but not limited to) robotics, flexible manufacturing, logistics and people management. The aim of this PhD scholarship is to investigate and reinforce the applicability of this kind of technique considering the whole spectrum of domains that recently emerged from the AIPlan4EU (aiplan4eu-project.eu) project. The candidate will research innovative approaches and algorithms to improve the performance, usability and/or relevance of planning and scheduling techniques deployed in diverse scenarios, having the unique possibility to work and experiment with real-world scenarios of planning already deployed by the project.

Planning - devising a strategy to achieve a desired objective - is one of the basic forms of intelligence, with applications in autonomous robotics, logistics, flexible production, and many other fields. Historically, planning research has followed a general-purpose framework: a generic engine searches for the strategy by reasoning on the problem statement. Despite substantial progress in recent years, domain-independent planning still suffers from scalability issues and fails to deal with real-word problems. The alternative is to devise ad-hoc, domain-specific solutions that, although efficient, are costly to develop, rigid to maintain, and often inapplicable in non-nominal situations.

The PhD student will study the foundations of an innovative approach to Planning that will be domain-independent and efficient at the same time. The idea is to adopt a framework based on Reinforcement Learning, where a domain-independent planner is specialized with respect to the domain at hand. This research project will advance the state of the art in planning beyond the “efficiency vs flexibility” dilemma and provide effective techniques to be validated on real-world use-cases.

Hybrid systems are formal models combining discrete and continuous-time dynamic behaviors. They can be found in various applications such as robotics, control systems, cyber-physical systems, and transportation systems. Formal methods for hybrid systems provide a powerful set of techniques for designing, analyzing, and verifying the behavior of complex systems that exhibit both continuous and discrete behaviors. These techniques can be used to ensure the correctness and safety of the system and to detect design flaws and bugs early in the development cycle. This project will investigate new formal methods to prove properties of hybrid systems integrating model checking, automated theorem, and numerical analysis for control theory. Different aspects of hybrid systems will be considered including temporal properties, diagnosability and epistemic properties, reliability and robustness to faults. Compositional reasoning and proof synthesis will be also considered. The new methods will be implemented and evaluated on industrial benchmarks derived by industrial collaboration of FBK in various application domains such as space, avionics, autonomotive, railyways, and energy.

With an increasing availability and need of point clouds and 3D urban models, the inclusion of semantic information is becoming more and more important, in order to facilitate the usage and exploitation of such data. Traditional deep learning methods applied to 3D geospatial data suffer of generalisation, adaptation and explainability. Data annotation is also a major bottleneck, being time consuming and prone to errors.

The research topic should work with photogrammetric, RGB-D and LiDAR 3D data and:
(i) investigate self-supervised and unsupervised 3D classification methods, including few-shot or zero-shot learning
(ii) design models that can better adapt and generalise among scenarios
(iii) make 3D semantic segmentation results more explainable.

This research position calls for a highly motivated and skilled researcher who possesses a good combination of computer science, AI and geomatics knowledge.

This research aims to develop an innovative framework that combines optical sensors (e.g. cameras and LiDAR) and Internet of Things (IoT) technologies for real-time monitoring of civil infrastructures (e.g. bridges, buildings, dams, etc.). This research position calls for a highly motivated and skilled researcher who possesses a good combination of software expertise, hardware integration knowledge and fast prototyping capabilities. Therefore the successful candidate is supposed to have a strong ability to bridge software and hardware components, along with the agility to rapidly prototype innovative, reliable and replicable solutions. The interdisciplinary Phd is expected to provide a secure, replicable and reliable framework whose findings will be applicable to various industries, including transportation, construction, mining, etc., and could have significant implications for public safety and economic development.

Conversational agents are experiencing a surge in interest given the continuous release of new models and the ever evolving scenario of NLG. Still, the actual focus is mainly on model size, training data size and prompt engineering. The interaction of these elements with related aspects, such as decoding strategies, knowledge guided generation, data quality, knowledge distillation -just to mention a few- can help in improving the models, especially for better factuality, reducing hallucination and increasing coherence among dialogue turns. The goal of this PhD Thesis is to overcome the shortcomings of present large language models by incorporating novel strategies for better generation.

In spite of the recent progress in speech technologies, processing and understanding conversational spontaneous speech is still an open issue, in particular in presence of challenging acoustic conditions as those posed by dinner party scenarios. Although enormous progresses have been made recently in a variety of speech processing tasks (such as speech enhancement, speech separation, speech recognition, spoken language understanding), targeting also multi-speaker speech recognition, a unified established solution is still far from being available. Moreover, the computational complexity of current approaches is extremely high, making an actual deployment in low-end or IoT devices not feasible in practice.

The candidate will advance the current state-of-the-art in speech processing (in particular for separation, enhancement and recognition) towards developing a unified solution, possibly based on self-supervised or unsupervised approaches, for automatic speech recognition in dinner party scenarios, as those considered in the CHiME challenges (https://arxiv.org/abs/2306.13734).

Industrial systems are reaching an unprecedented degree of complexity. The process of designing a complex system is expensive, time consuming and error-prone. Moreover, the design process has to guarantee not only the functional correctness of the implemented system, but also its dependability and resilience with respect to run-time faults. Hence, the design process must characterize the likelihood of faults, mitigate possible failures, and assess the effectiveness of the adopted mitigation measures.

Formal methods have been increasingly used over the last decades to deal with the shortcomings of designing a complex system. Formal methods are based on the adoption of a formal, mathematical model of the system, shared between all actors involved in the system design, and on a tool-supported methodology to aid all the steps of the design, from the definition of the architecture down to the final implementation in HW and SW. Formal methods include technologies such as model checking, an automatic technique to symbolically and exhaustively analyze all possible executions of the system in the formal model, in order to detect design flaws as early as possible. Model checking techniques have been recently extended to assess the safety and dependability characteristics of the design, and for system certification.

The objective of this study is to advance the state-of-the-art in system design using formal methods. This includes adapting and extending the system design methodology, investigating improved versions of state-of-the-art routines for verification and safety assessment of complex systems, and developing novel extensions to address open problems. Examples of such extensions include novel techniques for contract-based design and contract-based safety assessment, advanced techniques for formal verification based on compositional reasoning, the analysis of the timing aspects of fault propagation, the characterization of transient and sporadic faults, the analysis of the effectiveness of fault mitigation measures in presence of complex fault patterns, and the modeling of analysis of systems with continuous and hybrid dynamics.

This study will exploit the challenges and benchmarks defined in various industrial projects carried out at FBK.

Clinical neuroscience is playing a key role in the understanding of the brain with data of pathological alterations. The detection of anomalies in the brain structure and function is a crucial step not only for diagnosis and prognosis but also to decode the connectome of the human brain. Data driven approaches are providing promising results to characterize the patterns of the healthy brain. The challenge is to disentangle the intrinsic interindividual differences in the brain structure and function with respect to alterations related to cognitive impairment.

The research objective is to investigate the most innovative techniques of Artificial Intelligence, such as geometric deep learning, to translate the knowledge of connectivity structures from a healthy population to the individual patients of a clinical study. The ultimate goal is the development of computational methods to support the detection of altered structures in the connectome affected by brain disorders.

Clinical neuroscience is playing a key role in the understanding of the brain with data of pathological alterations. The detection of anomalies in the brain structure and function is a crucial step not only for diagnosis and prognosis but also to decode the connectome of the human brain. Data driven approaches are providing promising results to characterize the patterns of the healthy brain. The challenge is to disentangle the intrinsic interindividual differences in the brain structure and function with respect to alterations related to cognitive impairment.

The research objective is to investigate the most innovative techniques of Artificial Intelligence, such as geometric deep learning, to translate the knowledge of connectivity structures from a healthy population to the individual patients of a clinical study. The ultimate goal is the development of computational methods to support the detection of altered structures in the connectome affected by brain disorders.

This doctoral fellowship, co-funded by the Italian-German Historical Institute of the Bruno Kessler Foundation of Trento, is open to graduates in historical-archival and philological-literary disciplines and is aimed at the study of the following topics, with particular emphasis on the chronological span of the pre-modern era (from the 15th to the 19th centuries): the history of political and religious power; the history of communication and the media; the history of institutions; environmental history; urban history and the history of mobility.
Preference will be given to research projects and candidates that address the above topics from the perspective of European comparative history.

As we gather increasing volumes of data from next-generation sequencing and multi-omics technologies, the potential to develop transfer learning and neural networks applicable to diverse fields expands. A central challenge that remains is the creation of embeddings capable of representing information in a biologically meaningful space. The engineering and characterization of a model capable of representing this information in a comprehensive and robust manner could revolutionize the field with an impact comparable to what word2vec did for natural language processing.

The prospective PhD candidate is expected to work on the development of a model capable of recapitulating the entire biology of selected organisms, characterizing the manifold in terms of gene ontologies and pathway enrichments. The resulting encoding capabilities should be usable by the scientific community at large, and result in significant advancements in the field such as images to RNA inference models and vice versa.

Through this endeavor, we aim to make a significant stride in unifying diverse data types under a shared, biologically meaningful representational framework. This study will not only advance our understanding of biological systems but also serve as a herald for a new wave of AI applications in bioinformatics.

With the aging of global population, the prevalence of neurodegenerative diseases (among
which Alzheimer Disease, Parkinson’s Disease and Multiple Sclerosis) is rapidly increasing, representing a challenge for the sustainability of the healthcare systems in the forthcoming decades. Machine learning represents a promising tool to build models for the prediction of disease course and the targeting of therapeutic and care strategies. Nevertheless, traditional AI tools require lots of curated data to provide robust and reliable predictions, which might be a challenge when dealing with longitudinal clinical data. Indeed, clinical data are challenging to obtain due to privacy concerns and, whenever available, they might present many missing values, rendering them unsuitable for the training of AI algorithms.
The objectives of this research are two fold. The first is to build predictive models of the
development of several neurodegenerative diseases to assist clinicians in decision making.
The second objective is to investigate methods to mitigate data scarcity challenges. These
methods will exploit similarities between these diseases and will include a range of approaches such as transfer learning and generative models. Each of these approaches will
then be evaluated in terms of predictive performance in a set of clinically relevant outcomes.

Space systems have reached an unprecedented degree of complexity. The design process has to guarantee not only the functional correctness of the implemented system, but also its dependability and resilience with respect to run-time faults. Hence, the design process must characterize the likelihood of faults, mitigate possible failures, and assess the effectiveness of the adopted mitigation measures.

Formal methods have been increasingly used over the last decades to deal with the shortcomings of designing complex systems, in different domains. Formal methods are based on the adoption of a formal, mathematical model of the system, shared between all actors involved in the system design, and on a tool-supported methodology to aid all the steps of the design, from the definition of the architecture down to the final implementation in HW and SW.

The objective of this study is to advance the state-of-the-art in space system design using formal methods. In particular, it will investigate new techniques for model-based system and software engineering, to support the design, mission preparation and operations of space systems. The potential research directions include fault detection, isolation, and recovery for satellites; system level diagnosis and diagnosability based on telemetry; digital twins for satellites. Topics to be investigated include techniques for contract-based design and contract-based safety assessment, advanced verification techniques based on compositional reasoning, the analysis of the timing aspects of fault propagation, the characterization of transient and sporadic faults, the analysis of the effectiveness of fault mitigation measures in presence of complex fault patterns, and the modeling and analysis of systems with continuous and hybrid dynamics.

The developed techniques will be implemented and evaluated using tools for system-software engineering such as the COMPASS tool and the COMPASTA tool, based on the TASTE tool chain.

Water phase changes (evaporation, condensation, freezing) are ubiquitous phenomena of great importance for living beings and in engineering applications. The structure (micro and nano) and chemistry of surfaces control the kinetics and dynamics of these transitions. Plants, for example, offer numerous examples of self-cleaning, antifreeze and water-collecting1 properties developed over millions of years of evolution. Engineered anti-frosting surfaces find applications in aerospace (ice accretion on aircrafts), heat exchangers (refrigerators), wind turbines and power lines. Structured surfaces that increase evaporation and condensation efficiency are a challenge for Loop Heat Pipes (LHP) and Vapour Chambers that cool electronics on space stations (in microgravity conditions) or in the electronic devices we use on a daily basis. Surfaces that can efficiently collect dew and fog provide a source of water in arid environments and can improve the water recovery system of space stations.

This project will extend the studies carried out during the previous PhD scholarship (within cycle 34, in collaboration with FBK) which focused on anti-frosting and water-harvesting surfaces. The research activity will concern the theoretical study, fabrication, characterisation and experimentation of micro- and nanostructured surfaces with applications in aerospace and energy efficiency. In particular, phenomena of spontaneous jumps of condensation droplets on hydrophobic surfaces, distant coalescence on hydrophilic surfaces and freezing of droplets will be studied. Fabrication techniques may range from micro- and nanolithography, focused io beam, etching, chemical deposition processes, and polymer moulding. The expected outputs are patents and publications on high impact journals in the field.

The European Space Agency (ESA) mission JUpiter ICy moons Explorer (JUICE) to the Jovian system has  been launched in April 2023, and the development of the ESA EnVision mission to Venus will continue. Both missions carry on board a radar sounder instrument for subsurface sensing of planetary bodies. In the context of these two projects we are looking for candidates willing to design methodologies for sub-surface radar image processing and analysis. The outcome of this activity will contribute in improving the understanding of subsurface structures, and their correlation to planetary body history and climate.

The candidate will be requested to design and develop novel methodologies based on machine learning, deep learning, pattern recognition and artificial intelligence for information extraction, classification, target detection, noise reduction and change detection in radar and radar sounder images.

Besides the requirements established by the rules of the ICT school, preferential characteristics for candidates for this scholarship are:

• master degree in Electrical Engineering, Communication Engineering, Computer Science, Mathematics or equivalents;

• knowledge in pattern recognition, deep learning, image/signal processing, statistic/remote sensing/radar.

This grant is funded by project ASI JUICE-RIME-E - “Missione JUICE - Attività dei team scientifici dei payload per lancio, commissioning, operazioni e analisi dati” — CUP F83C23000070005.

One of the pillars of healthcare digital transformation focuses on the integration of AI-based solutions within the clinician-patient relationships with the aim of monitoring and/or supporting them toward the achievement of healthy functional status.

Examples of these systems are (i) virtual coaches to support remote monitoring and recommendations for patients affected by stress or nutritional chronic diseases; (ii) telehealth solutions to enhance the care capabilities of health organizations; and, (iii) tools to orchestrate care pathways involving, besides patients, multiple clinical actors.

This Ph.D. works within this context with the aim of exploiting behavior change intervention (BCI) strategies to design novel persuasive frameworks nurturing AI-based systems to trigger the implementation of the next-generation virtual digital assistants.

The area of intervention is very broad since the research areas involved are, for example, human-computer interaction, psychology, computation linguistics, knowledge management, and pervasive computing.

For this reason, the Ph.D. candidate will have the opportunity to explore the virtual digital assistants' domain in order to analyze current open challenges, decide which ones to address, and strategies she/he will use to tackle such challenges.

The Ph.D. candidate will also have the opportunity to validate her/his solutions in several real-world settings with the aim of validating her/his work in practice.

The aim of this Ph.D project is to explore innovative, technology-based approaches to facilitate playful education for inclusive, collaborative and student-centered learning environments. The project aims to develop semi-automatic and adaptive educational paths, as well as personalized support systems that benefit both educators and learners. This will involve utilizing various technologies and methods, such as end-user programming or multimodal, competence-based learning, to co-create learning activities that consider specific needs and preferences.
The ideal candidate will have a background in Computer Science, Psychology or Cognitive Science, with knowledge of data science, artificial intelligence, and educational theories. Experience in designing interactive digital technologies, conducting experimental and in-the-wild studies, and promoting inclusive, collaborative learning environments would be advantageous and should be gained during the Ph.D training.

This project will develop an innovative platform in thin film lithium niobate for performing quantum optics experiments with integrated devices. It will focus on the development of innovative sources for single photons and squeezed vacuum based on periodically poled waveguides and reconfigurable circuits, and on the integration of superconducting single photon detectors on-chip. This project is part of an on-going collaboration between the Department of Industrial Engineering at the University of Trento, the Bruno Kessler Foundation (FBK), the Italian Institute for Nuclear Physics (INFN), and the National Research Council (CNR).

The fundamental objective of the PhD activity is the development of prototype cells of ionic conductive cells. It is possible to divide the planned activities into two main activities, closely related to each other and aimed at optimizing the devices and their performance. First activity will be performed in strong collaboration with UniTN-DII, concerning the production of materials and the realization of the prototype cells. Reference will be made to planar architectures where, initially, the compositions will refer to an electrolyte based on zirconia or Ceria suitably doped, hydrogen side electrode based on zirconia/nickel/copper or perovskites such as ferrites, manganites or cobaltites, the latter to be also use for the oxygen side electrode; compositions and structures will be identified above all on the basis of the expected operating temperatures and suitably modified in order to optimize performance, always keeping in mind the production process which will have to allow for the achievement of multilayer planar structures "alla ceramic". The prototype cells will be made using colloidal technologies (tape casting, screen printing, digital printing) and subsequent co-sintering to then be characterized from a chemical-physical and structural point of view to be able to correlate the performances as identified in the subsequent activity. The second activity provides for the physicochemical, structural, and electrochemical characterization in electrolysis and fuel cell mode of the cells, in collaboration with FBK.

The candidate will work in the Integrated and Quantum Optics group, FBK, where he will develop technologies for integrated photonics.

The technological platforms on which the candidate will operate will be mainly Lithium Niobate On Insulator (LNOI) and silicon. LNOI is one of the most interesting materials for integrated photonics due to the wide transparency window and its non-linear optical properties. The strong and fast electro-optic responsivity of Lithium Niobate will be exploited to design fast modulators and highly efficient correlated photon sources.

The candidate will develop techniques for thin films of LNOI processing aimed at the realization of such photonic devices on the platform. The candidate will exploit the technological platform to design optical devices and components, that he/she will fabricate and characterize in the laboratories of FBK and UniTN.

Redox flow batteries (RFBs) are a promising technology for large scale energy storage. In RFBs power and energy are decoupled: the former depends mainly on the size of the stack while the latter on the size tanks containing the redox active species. This feature makes RFBs ideal for economical, large-scale energy storage. However, cost reductions are needed for a widespread diffusion of this technology. The required cost reductions involve two main components of the system: the electrolytes and the stack. Both need to be optimized for enabling a large-scale deployment of RFBs. RFBs are a complex system so a trade-off between cost and performances must be found for both electrolytes and stack.

This work will focus on developing and validating the use of low-cost earth-abundant redox active materials (e.g., Al, Zn, Fe, Cu, Mn, I, S) in redox flow batteries analyzing different possible solutions for redox couples and electrolytes. The doctoral thesis will start with the development of an integrated approach for the evaluation of different electrolyte solutions (combining different redox active materials, supporting electrolytes and solvents) assessing material cost and availability (with a special focus on EU-sourced materials), safety, expected performances, probability of upscaling issues (e.g., cross-over, pumping losses, etc) and other parameters in other to select the most promising solutions. This pre-selection will be refined by integrating experimental data obtained from the electrochemical and physical characterization of the electrolytes. The use of solid boosters and/or complexing agents to improve the energy density will be a fundamental step to be evaluated experimentally. The selected best candidates will be then tested in lab-scale redox flow cells customized for the specific chemistry. Membrane/separator selection or development will also be addressed during this phase. Finally, issues related to the upscaling of the technology will also be analyzed in order to assess the real potential impact of the proposed innovative solutions.

This research aims to investigate the integration of distributed learning paradigms with robotics, IoT, and edge computing in the context of digital agriculture. It will explore the potential benefits of using distributed learning approaches, such as federated learning, continual learning, and reinforcement learning to improve crop yields and reduce costs and environmental impact by enhancing the overall efficiency of agricultural operations, also including active sensing, autonomous navigation, and decision-making. The research will explore IoT and edge computing to support the collection, processing, and analysis of data from distributed sensors in the field. The study will investigate the technical and economic factors that influence the adoption and implementation of these technologies and frameworks in agriculture, including issues related to scalability, resource constraints, and interoperability. Finally, the research will also explore the potential of combining these technologies and frameworks to create new opportunities for innovation and collaboration in digital agriculture.

Neutrons are a very particular particle, with wide and unique applications for both fundamental studies and as a probe. Thermal neutrons are detected by nuclear interactions, in which their energy information is lost. This project, following on from a proof of concept and a statistical investigation of using the cross section variation of the interaction with energy to extract the neutron energy, will use these in combination with AI/machine learning to produce a practical algorithm to determining the neutron energy and the limits of the technique with multilayer detectors. It will also look at the experimental implement and verification of this. The outcome is a determination of the thermal neutron’s energy and an algorithm to do so in a realisable detector, which is very much the holy grail of neutron detection.

Nowadays, microelectromechanical systems (MEMS) represent a well-established class of miniaturized devices integrating both mechanical and electrical components onto the same silicon board, fabricated together via photolithography-based process, which enables small device footprint, high sensing/actuating performances, low-power consumption, and mass production. MEMS devices, with particular reference to inertial sensors, are commonly applied in a variety of fields, ranging from consumer electronics to automotive industry, and have enabled novel functions and opportunities of the systems where they have been applied (e.g., screen rotation in mobile phones). However, they are still scarcely used in high-demanding applications, such as aerospace, because of insufficient sensitivity and stability, stemming from technological limitations affecting current fabrication strategies.

In this research project, novel technological solutions and mechanisms will be sought to develop MEMS inertial sensors based on Silicon-On-Insulator (SOI) wafer micromachining to enable high sensitivity and low noise, compared to standard devices. Different layouts will be developed and their mechanical and electro-mechanical performance compared through both analytical and numerical models, based on Finite Elements Method (FEM). The optimized layouts will be then fabricated and experimentally validated.

The research activity aims to develop novel catalyst layers for proton exchange membranes (PEMs) in water electrolysis. The objective is to enhance the efficiency and durability of PEM electrolyzers, which are key components in hydrogen production systems.
The expected outcome of this research is the development of advanced catalyst layers that exhibit improved electrochemical performance and stability. By designing catalyst materials with high activity, selectivity, and durability, researchers aim to enhance the efficiency of the electrolysis process, reduce the energy consumption, and extend the lifespan of the PEM electrolyzers.
These novel catalyst layers are expected to enable more efficient and cost-effective hydrogen production through water electrolysis. By optimizing the catalyst composition, structure, and interaction with the membranes, the research activity seeks to overcome the limitations of existing catalyst layers and contribute to the advancement of sustainable hydrogen production technologies.
In the production process of novel catalyst layers for proton exchange membranes (PEMs) in water electrolysis, various techniques are employed to achieve optimal morphology and composition. These techniques include deposition methods, such as physical vapor deposition (PVD) or electrochemical deposition (ECD), and synthesis methods like sol-gel or wet chemical methods.
The morphology of the catalyst layer plays a crucial role in determining its catalytic activity and stability. The student will focus on controlling factors such as particle size, shape, and distribution within the catalyst layer. These parameters directly influence the accessibility of reactants to the active sites, mass transport properties, and overall electrochemical performance.
In addition to morphology, the choice of support material is critical for the catalyst layer' performance. Support materials provide structural stability and serve as a platform for catalyst deposition. Common support materials include metal oxides, including titanium dioxide or cerium oxide and novel material as silicon carbide, and nitrides (vanadium, et.). The selection of the support material depends on factors such as its electrical conductivity, chemical compatibility with the catalyst and electrolyte (as pH and polarization), and durability under the operating conditions of the PEM electrolyzer.
By carefully controlling the productive process and optimizing the morphology and support material, the student aim to develop catalyst layers that exhibit high catalytic activity, stability, and efficient charge transfer.
The outcome of this research has the potential to significantly impact the field of water electrolysis by providing a pathway for the commercialization of PEM electrolyzers with enhanced performance and longevity. The development of novel catalyst layers can accelerate the adoption of hydrogen as a clean energy carrier, facilitating the transition to a more sustainable and

Sony Semiconductor Solutions Group (a branch of Sony Corporation) is a corporate group that conducts research and development, product planning, design, production, and sales of semiconductor-related products, centering on image sensors in the imaging field such as smartphones and digital cameras, and in the sensing field such as automotive, security, and industry areas. A new R&D centre has been opened in Trento and is offering a PhD scholarship jointly with Fondazione Bruno Kessler. The objective of the collaboration within this scholarship is to study neural network architectures suitable for the integration within next generation advanced CMOS image sensors and optical sensors targeting to address the following challenges: i) fully parallel connection and processing between the top-level pixel array and the bottom layers digital processing blocks, ii) high reuse of on-chip memory and computational resources, iii) low power operation.
The doctoral candidates will have to tackle the challenges with a multidisciplinary point of view, starting from the design and simulations of ideal neural network architectures implementation, developing modelling of sensor+NN, investigating design trade-offs between computational power and chip integrability.
The outcome of the research would be the result of a feasibility study and a proposed set of features/ hardware architecture which would allow the start of a development in this field. The intellectual property of the research results that will derive from the activities carried out by the doctoral student is owned by FBK and Sony.

Research project aimed at developing generative models for omics data with particular attention
to single cell RNA seq, focusing on metrics engineering to test data likelihood, enriching the
latent space with biological information, exploring transfer learning for model generalization,
implementing algorithms for trajectory estimation and bifurcation analysis, enriching differential
gene expression analysis with probabilistic models, and developing robust approaches for
personalized medicine using optimization and reinforcement learning algorithms. The proposed
models will be applied to various datasets, including viral and bacterial infections and cancer, in
order to validate their ability to recapitulate experimentally observed cellular cascades.

Progetto di ricerca finalizzato allo sviluppo di modelli generativi per dati omici con particolare
attenzione al single cell RNA seq, lavorando sull'ingegnerizzazione di metriche per validare la
verosimiglianza dei dati, sull'arricchimento dello spazio latente con informazioni biologiche,
l'esplorazione del transfer learning per la generalizzazione del modello, l'implementazione di
algoritmi per l'analisi di traiettorie cellulari, l'arricchimento dell'analisi differenziale di espressione
genica con modelli probabilistici e lo sviluppo di approcci robusti per la medicina personalizzata
utilizzando algoritmi di ottimizzazione e reinforcement learning applicati ai modelli generativi. I
modelli proposti saranno applicati a vari set di dati, comprese infezioni virali, batteriche e
cancro, al fine di validare la loro capacità di ricapitolare le cascate cellulari osservate
sperimentalmente.

Performance tests of Proton conductive ceramic cell for FC/EL particularly for Compression separation (from N2, cracking NH3, H2/H2O) with dedicate Durability test or AST on PCC

Research focus will be on the development and the validation of multiphysics-multiscale tools that will accelerate the optimal design and management of redox flow batteries, a technology particularly suited for long-duration energy storage. Flow batteries are a complex system and a multitude of physical phenomena involving different scales (from micro to system) need to be considered to forecast cell and battery system performance and optimize the design of each component.

Extracting information from clinical documents in a multilingual perspective.

Investigation of the effectiveness of cognitive-behavioral psychological interventions based on virtual reality and artificial intelligence.

The main goal of the research is to build an embodied system capable of interacting with an unknown environment and continuously learn about such an environment by exploiting commonsense knowledge. The experimental evaluation of the approach should involve both simulated environment and real environments. Aspects of sim2real transfer will also be considered.

One of the pillars of healthcare digital transformation focuses on the integration of AI-based dialogue solutions within the clinician-patient relationships with the aim of monitoring and/or supporting them toward the achievement of healthy functional status.
The advent of large language models (LLMs) changed the research paradigm in natural language processing (NLP) and it put the basis for new challenges.
Two of the most main issues of such LLMs are identified by the analysis of how they are ethical and sustainable.
By “ethical”, we mean their capability of both building models that are not biased (in a broader sense) and generating content that is controlled with respect to the task they are used for and the environment in which they are deployed.
Instead, as “sustainable”, we mean the possibility of training and fine-tuning LLMs with costs that may be afforded by any academic entity.
In this Ph.D., the candidate will investigate (i) how to design dialogue systems in a sustainable way by performing fine-tuning operations on existing LLMs; and, (ii) how to inject grounded knowledge into LLMs to enable ethical and bounded conversations between the AI-based system and the target users.

Thanks to the increasing digitalisation of contemporary organizations, event data about the execution of processes are continuously collected. Process intelligence aims at transforming these data into insights into how processes are executed in reality.

In spite of the reported effectiveness of state of the art approaches for process intelligence, an effective intelligent usage of execution data is still reported to be one of the key challenges of today's strategic management. Indeed, AI is becoming increasingly central to realise a view that shifts from descriptive and reactive predictive analytics to proactive prescriptive analytics, able to improve the impact, sustainability, and relevance of data-enabled decisions. Realising this vision requires addressing several foundational challenges: (i) the representational descriptive challenge, which arises from the multiperspective representation of knowledge for the specific domain containing time, resources, data objects, costs ...; and (ii) the predictive and proactive challenge, which demands the usage of data in an explainable manner and to realise recommender systems able to deal with the temporal dimension of data so as to support decision making. It also calls for an integrative approach that leverages the power and flexibility of Machine Learning techniques, combines the (implicit) knowledge contained in the data with the explicit (and often de facto or legally binding) rules governing the process behaviour, and resorts to different reasoning techniques. The PhD project will focus on the development of new AI based techniques able to deal with some of the challenges above and their application in several domains, e.g., the healthcare one.

Applications that rely on the most modern sensing devices and technologies and combine complex artificial intelligence tasks are now mainstream. It is sufficient to say, “OK-Google/Alexa/Siri switch on the heating system when the temperature is below 18° C” to appreciate the power of the IoT in combination with an Artificial Intelligence engine. However, the typical approach to enable intelligent applications is cloud-centric, meaning that the intelligence (a home assistant) is hosted in the cloud infrastructure, and the sensor data collected by some IoT devices (a microphone array and a temperature sensor) flow from the cyber-physical-system until reaching a remote endpoint to be processed. Finally, the correct command is transmitted to the IoT actuator (a radiator thermostat). Alternative approaches to this are possible, for instance, by considering a more dynamic and configurable intermediate layer placed between the IoT and the Cloud sides, usually dubbed as the Edge layer. Generally, a configurable edge layer reduces the required bandwidth and latency and improves users’ privacy. Moreover, if portions of the application intelligence could be hosted in this layer, the IoT device lifetime would be enlarged. However, reconfiguring and deploying an end-to-end processing flow that involves the three aforementioned architectural layers poses major challenges. Select a more efficient detection algorithm from a rich machine learning algorithms library and pushing the “deploy” button of an application dashboard to see the selected algorithm up and running more effectively (according to a given metric) on my smart home devices is still a dream, in most of the cases. Moreover, depending on the hardware capabilities, the application requirements in terms of bandwidth and latency, and the accuracy required for the machine learning task to execute, different end-to-end configurations are possible, all sub-optimal and possibly non-dominated in the Pareto meaning. The subject of this Ph.D. is to investigate and propose novel optimization and assessment methodologies to efficiently sample such a complex design space in target application sectors such as home, industry, manufacturing, farming, etc. The reference technological environment covers (but is not limited to) embedded device software engineering (micropython, mbed OS, C languages and dialects, etc.), machine learning frameworks deployable on tiny devices (tinyML, TensorFlow lite, etc.), edge-based frameworks (eclipse Kura, edgeX Fundry, etc.) and cloud-based IoT platforms and services with AI support.

The EU promotes the creation of data spaces in order to maximize the impact that gathered data can have on society and the environment. At the same time municipalities and regional governments push for new acquisitions of multimodal geospatial data on their territories (cities, forests, rural areas, etc.). However, the integration of such data at different resolution, radiometry and acquisition systems is one of the biggest obstacle for proper data exploitation. Therefore the goals of the proposed PhD are:
(i) to study, develop and validate innovative solutions to fuse geospatial data (such as 3D point clouds, multi/hyperspectral orthoimages, etc.) of the built/vegetated environment ad extract metric information;
(ii) to conduct research on novel and efficient algorithms for 3D data semantic segmentation and classification using integrative AI approaches that can effectively replace traditional hand-crafted methods to ultimately improve performance and interpretability, handle unbalanced classes, ease deployment and foster scalability;
(iii) to analyze, realize and demonstrate new methods to create digital twins of our environment using multimodal geospatial data.

Many application domains require the ability to automatically generate a suitable course of actions that will achieve the desired objectives. Notable examples include the control of truck fleets for logistic problems, the organization of activities of automated production sites, or the synthesis of the missions carried out by unmanned, autonomous robots. Planning and scheduling (P&S) are fundamental research topics in Artificial Intelligence, and increasing interest is being devoted to the problem of dealing with timing and resources. In fact, plans and schedules need to satisfy complex constraints in terms of timing and resource consumption, and must be optimal or quasi-optimal with respect to given cost functions. The Ph.D. activity will concentrate on the definition of an expressive, formal framework for planning with durative actions and continuous resource consumption, and on devising efficient algorithms for resource-optimal planning. The activity will explore the application of formal methods such as model checking for infinite-state transition systems, and Satisfiability and Optimization Modulo Theories, and will focus on practical problems emerging from the flexible manufacturing domain.

The advent of Industry 4.0 has made it possible to collect huge quantities of data on the operation ofcomplex systems and components, such as production plants, power stations, engines and bearings. Based on such information, deep learning techniques can be applied to assess the state of the equipment under observation, to detect if anomalous conditions have arised, and to predict the remaining useful lifetime, so that suitable maintenance actions can be planned. Unfortunately, data driven approaches often require very expensive training sessions, and may have problems in learning very rare conditions such as faults. Interestingly, the systems under inspection often come with substantial background knowledge on the structure of the design, the operation conditions, and the typical malfunctions. The goal of this PhD thesis is to empower machine learning algorithms to exploit such background knowledge, thus achieving higher levels of accuracy with less training data.

Runtime verification is a light weight verification technique based on the analysis of system logs. A key factor is that the internal state of the system is not observable, but partial knowledge on its behaviour may be available. The thesis will investigate the use of temporal epistemic logics (i.e. logics of knowledge and believe over time) to specify and verify hyperproperties for runtime verification. Different logical aspects, like distributed knowledge and common knowledge, and the communication between reasoning agents, will be used to model hierarchical architectures for fault detection and identification, and for prognosis. Techniques for planning in belief space will be used for the design of fault reconfiguration policies.

Many artifacts in the development process (requirements, specifications, code) tend to become legacy, hard to understand and to modify. This results in lack of reuse and additional development costs. A reverse engineering activity is necessary to understand what the system is doing. Goal of the thesis is to provide automated techniques to analyse the inherent behavior of legacy artifacts, extract interface specifications, and to support re-engineering activities. The thesis will combine techniques from language learning, applicable to black-box artifacts, and formal techniques for the automated construction of abstractions in the form of extended finite state machines.

Simulation tools are fundamental to predict the evolution of pandemic, and to assess the quality of counter-measures, e.g. the effect of travel restrictions on the spread of the coronavirus. However, they come with two fundamental requirements. The first is the need for a fast reconfiguration of the simulation, in order to be able to describe the mutating scenarios of the pandemics. The second is the ability to produce correct and explainable results, so that they can be trusted and independently validated. The topic of this research is to devise a model-based approach that is able to represent at a high-level the features of a generic pandemic, from which an efficient simulator can be produced. Using formal methods, the results of the simulation are guaranteed to be correct by construction, with proofs that can be properly visualized and independently checked. The activity will be carried out as a collaboration of the Center for Health Emergencies (https://www.fbk.eu/it/health-emergencies/), that played a major role during the ongoing pandemics, and the Center for Digital Industry (https://dicenter.fbk.eu/), a leading centers in model-based design.

Health data and machine learning are increasingly used to tackle fundamental challenges in medicine, ranging from providing a better understanding of disease trajectories, improving care services or devising new policies. As AI-supported decision-making makes its way into the clinical practice, of particular concern is presence of representation bias, where groups from particular socio-economic, racial, ethnicity and religious backgrounds might be underrepresented in the health data sets, used to devise AI predictive models. Consequently, the resulting decisions might be discriminatory, with the potential to perpetuate existing health inequities.
In response, several approaches have been developed to address health data poverty, including data augmentation. Data augmentation methods are used to generate synthetic data from underrepresented groups that closely resemble real data to mitigate representation bias.
Machine learning community has developed various approaches to synthetic data generation, ranging from simple resampling methods, such as SMOTE, up to recent methods based on artificial neural networks, using generative adversarial networks (GAN). These methods have been proven to work very well in imaging data, however several research challenges remain in generating longitudinal, multivariate data. Furthermore, none of the existing approaches are designed to specifically target the challenge of health data poverty.
Therefore, the focus of the PhD work will be on devising novel GAN architectures that can generate high-dimensional, longitudinal clinical data to mitigate representation bias. Several datasets will be considered in this work containing patient data from underrepresented groups, such as US-based critical care datasets (MIMIC IV and eICU-CRD) as well as European based datasets, including National Multimorbidity Resource provided by Health Data Research UK (HDRUK), Connected Bradford and UK Biobank. The resulting methods will be data agnostic and applicable to a wide range of datasets.

Medical imaging techniques, such as computed tomography (CT) and positron emission tomography (PET), play a crucial role in early detection, diagnosis, intervention, and follow-up of patients with lung cancer, one of the leading causes of cancer-related mortality worldwide. In today’s clinical practice, only a few quantitative PET/CT metrics are being directly used to describe tumors, such as CT-based largest diameters of nodules, and metrics derived from PET’s standardized uptake value. “Radiomics” is a field of research that emerged in recent years, in which medical images are converted into minable quantitative data that can be further analyzed by artificial intelligence (AI) algorithms or conventional statistics, deriving a shortlist of radiomics features representing non-invasive biomarkers for lung cancer risk or survival prediction. Typically, different AI models can be built, either using features extracted from single-modality images (e.g. CT only or PET only) or using combinations of features extracted from multimodal images through different levels of fusion.
A different and complementary approach leverages deep learning (DL) to automatically extract “deep features” from whole PET/CT images, without the need for explicit feature definition. However, DL models typically lack intuitive interpretability due to their “black box” nature.
The proposed PhD project aims at combining radiomics and DL on an original PET/CT data set (APSS Trento) to predict a clinical outcome such as patient survival. Radiomics features will be input to a predictive model (such as Random Forest or Logistic Regression) to investigate the association with the clinical phenotype(s) of interest. In parallel, an approach based on DL will be implemented, exploiting the whole medical images to learn completely data-driven “deep features”. To overcome the limited interpretability of DL models, particular care will be taken to supplement the developed models with post-hoc explainability techniques. Lastly, both radiomics and DL features will be combined into an integrated model predicting the association with clinical phenotypes. One of the crucial methodological aspects will be the adoption of techniques to avoid overfitting and guarantee reproducible results: for example, dimensionality reduction (PCA) and feature selection (Ridge, Lasso), as well as cross-validation, and, particularly for DL, data augmentation and transfer learning.

Si tratterà di studiare temi di storia ambientale in un periodo compreso tra l’epoca moderna e quella contemporanea. In particolare affronterà lo studio del rapporto tra natura e società con un focus sull’area alpina, intesa in senso lato e anche in prospettiva comparata.

Nowadays, digital identities are employed by the majority of European governments and private enterprises to provide a wide range of services, from secure access to social networks to online banking. As the Digital 2023 global overview report shows, the number of digital identities is growing: we have 4.76 billion social media users and spend trillions of dollars on e-commerce.
Digital identity is therefore a key ingredient for securing new IT systems and digital infrastructures such as those based on zero trust. For these reasons, the secure deployment of digital identity solutions is a mandatory prerequisite for building trust in digital ecosystems and is an obligation shared by security practitioners and consumers.
The research work to be conducted in the thesis aims to develop a new approach for automated security, privacy, and risk management in the design, development, and maintenance of digital identity solutions. The challenge is to deal with the multiple dimensions of the design space as a continuum in which specifications are analyzed both in isolation and as refinements of each other.
The approach should take into account the specific security and privacy issues of each phase and, at the same time, consider the interdependencies among the design and implementation choices performed in the various phases, bridging the gap among them.
The resulting approach should be automated, auditable, provide actionable hints to reduce risk, and be easy to integrate into the wide range of services and applications that arise in the plethora of use case scenarios resulting from the pressure of digital transformation. This activity includes:
- Analysis of state-of-the-art identity management solutions and their security issues.
- Identification of relevant use cases.
- Specification of a (semi-)automatic approach for security and risk management of digital identity solutions.
- Implementation of the approach on a tool and experimental evaluation on real-world use cases.

References:
M. Pernpruner, R. Carbone, G. Sciarretta and S. Ranise. An Automated Multi-Layered Methodology to Assist the Secure and Risk-Aware Design of Multi-Factor Authentication Protocols. Submitted to IEEE Transactions on Dependable and Secure Computing (TDSC).
G. Sciarretta, R. Carbone, S. Ranise, L. Viganò. Formal analysis of mobile multi-factor authentication with single sign-on login. ACM Transactions on Privacy and Security (TOPS) 23 (3), 1-37.
M. Pernpruner, R. Carbone, S. Ranise, G. Sciarretta. The Good, the Bad and the (Not So) Ugly of Out-of-Band Authentication with eID Cards and Push Notifications: Design, Formal and Risk Analysis. In Proceedings of the Tenth ACM Conference on Data and Application Security and Privacy (pp.223-234), 2020.
A. Sharif, M. Ranzi, R. Carbone, G. Sciarretta, F. A. Marino, S. Ranise. The eIDAS Regulation: a Survey of Technological Trends for European Electronic Identity Schemes. MDPI Journal of Applied Science (APPLSCI), 2022.

Industrial systems are reaching an unprecedented degree of complexity. The process of designing a complex system is expensive, time consuming and error-prone. Moreover, the design process has to guarantee not only the functional correctness of the implemented system, but also its dependability and resilience with respect to run-time faults. Hence, the design process must characterize the likelihood of faults, mitigate possible failures, and assess the effectiveness of the adopted mitigation measures.

Formal methods have been increasingly used over the last decades to deal with the shortcomings of designing a complex system. Formal methods are based on the adoption of a formal, mathematical model of the system, shared between all actors involved in the system design, and on a tool-supported methodology to aid all the steps of the design, from the definition of the architecture down to the final implementation in HW and SW. Formal methods include technologies such as model checking, an automatic technique to symbolically and exhaustively analyze all possible executions of the system in the formal model, in order to detect design flaws as early as possible. Model checking techniques have been recently extended to assess the safety and dependability characteristics of the design, and for system certification.

The objective of this study is to advance the state-of-the-art in system design using formal methods. This includes adapting and extending the system design methodology, investigating improved versions of state-of-the-art routines for verification and safety assessment of complex systems, and developing novel extensions to address open problems. Examples of such extensions include novel techniques for contract-based design and contract-based safety assessment, advanced techniques for formal verification based on compositional reasoning, the analysis of the timing aspects of fault propagation, the characterization of transient and sporadic faults, the analysis of the effectiveness of fault mitigation measures in presence of complex fault patterns, and the modeling of analysis of systems with continuous and hybrid dynamics.

This study will exploit the challenges and benchmarks defined in various industrial projects carried out at FBK.

CMOS Application Specific Integrated Circuits (ASIC) coupled with silicon radiation detectors provide the best performance in terms of noise, power consumption and compactness in a wide range of applications. Some of them include: biomedical spectroscopy or tomography, Earth-based or space telescope for monitoring x or gamma ray events coming from deep space, physic experiments, security scanning, and cultural heritage monitoring.
The system detects the photons that are converted in a voltage/current signal processable by dedicated mixed-signal electronics implementing different on-chip functionalities, such as photon counting, energy spectrum, energy-based discrimination and imaging. In this field, today's main challenges are the achievement of low electronic noise and power consumption, combined with high photon rates working operations.
The goal of this project is to develop a CMOS ASIC specifically designed for silicon radiation detector as, for example, Silicon Drift Detector (SDD), Low Gain Avalanche Diode (LGAD) and Silicon Photon Multiplier (SiPM), focusing on low-noise, high-rate and low power consumption trade-off.
The student will interact with experts in the fields of radiation sensors and analog/mixed signal integrated circuit design, gaining a unique combination of background knowledge. The expected outcome is the realization of state-of-the-art ASIC for radiation detectors and their validation in a real use-case scenario.

The PhD candidate will focus on the development and microfabrication of superconducting quantum devices, such as devices based on Josephson junctions and/or devices based on superconducting high-kinetic-inductance films. The candidate will be involved in the design and simulations and in the microfabrication of the devices, exploiting the cleanroom facilities at Fondazione Bruno Kessler. The cleanroom processes for the microfabrication of the single circuitry elements, such as Josephson junctions, microwave resonators and high-kinetic-inductance films, will be optimized. Subsequently, the optimized processes will be exploited for the fabrication of the final quantum devices, which will be characterized at millikelvin temperature by means of dilution refrigerators. The results of the measurements and of the corresponding data analysis will serve to further optimize the device in order to improve the performance. The optimized devices (Josephson junctions-based devices and/or Kinetic Inductance Travelling Wave Parametric Amplifiers) will be exploited to investigate fundamental physics processes, such as microwave squeezing and generation of entangled microwave photons. The produced devices will be finally combined to create a complete read-out chain for multiqubit systems.

This PhD project will tackle fundamental challenges related to Integrative AI for Natural Language Understanding (NLU), through the combination and integration of state of the art neural models according to the following dimensions:
- neural models able to integrate data and knowledge from heterogeneous sources;
- neural models that are well integrated with societal needs;
- neural models that can be adapted into realistic and rich communicative contexts;
- neural models that can be combined to obtain comprehensive NLU models.

The research conducted during the PhD will be aimed to:

- improve methods for assessing the risks of mosquito-borne infections;
- quantify the effectiveness of alternative interventions, both existing and prospective, in reducing mosquito populations and/or disease transmission;
- support public health decision makers in preparedness against emerging and re-emerging mosquito-borne infectious threats;
- improve scientific knowledge through the publication of high-impact papers.

A large wealth of data, collected by Italian and international leaders in entomological and public health surveillance, will be made available to support and calibrate next-generation models of mosquito-borne infectious diseases.

Optical and structural characterisation of stress-controlled films deposited on monocrystalline silicon wafers and defined by unconventional photolithographic processes.

The focus of the PhD grant will be on the extraction of information from clinical documents in a multilingual perspective, with a specific emphasis on languages with limited resources.

In the last decade there has been a growing interest for virtual coaching interventions delivered by means of mobile applications and personal assistants to support self-care of patients, including those coping with cancer. Despite the fact that the validity of virtual coaching treatments has been proved repeatedly by previous research, the design of effective behavioral intervention technologies for patient care and well-being remains a challenge.  The aim of the PhD project is to investigate key features of smart coaching solutions for patient care and well-being interventions that are engaging to use by patients and produce effective outcomes from a clinical perspective. The ideal candidate will be strongly motivated in developing design skills in the field of behavioral intervention technologies and conversational agents for applications in healthcare. The PhD position is offered in co-tutoring between the DHLab unit of FBK and the Department of Psychology of the University of Padova.

The focus of the PhD grant will be on the extraction of information from clinical documents in a multilingual perspective, with a specific emphasis on languages with limited resources.

The research theme of this doctoral position will concern automated learning and inference by means of hybrid AI models, including their application to the interpretation of multimodal data.

Automated Planning is the problem of synthesizing courses of actions guaranteed to achieve a desired objective, given a formal model of the system being controlled. A class of problems particularly interesting for applications is temporal planning (also called planning and scheduling) where the discrete decisions of "what to do" are coupled with the problem of scheduling (deciding "when to do"). Planning and scheduling techniques are important in several application domains such as flexible manufacturing and robotics. Unfortunately, these techniques suffer from scalability issues and are often unable to cope with the complexity of real-word scenarios, despite the significant advances in the field.

Recently, efforts such as Deepmind AlphaZero and OpenAI Five hit the headlines, with groundbreaking advancements in the field of reinforcement learning. These techniques are able to automatically learn policies to decide what to do in order to achieve a desired goal. However, they offer no formal guarantee and are not model-based. The research objective of this PhD scholarship is to investigate techniques that combine the formal guarantees offered by automated planning and scheduling with the performance and self-improving capabilities offered by recent advances in deep reinforcement learning to construct adaptive planners that can learn strategies capable of solving problems in a specific application scenario and improve their performance (in terms of both speed and quality) over time.

AI components are more and more used in safety-critical systems in different application domains such as automotive or space. In particular, the increased availability of sensor data gives the opportunity to increase the autonomy these systems with advanced perception, optimized control, and efficient fault detection and recovery. The validation, verification, and safety assurance of AI components in these systems are therefore of paramount importance. However, the uncertainty of Machine Learning (ML) algorithms poses hard challenges for traditional approaches. In this PhD project, we aim at investigating new model-based design techniques to ensure the safe usage of AI/ML components. We will explore the definition of new formal models to represent the uncertainty of the ML models and the related errors, as well as formal verification techniques for the evaluation of the reliability of the system with AI components, and will design and evaluate architectural schemas in specific application scenarios.

While the task of automatically recognising hateful content online has been extensively explored in the last years within the NLP community, what is the best strategy to respond to such messages has only recently entered the research agenda. One of the main issues related to this task is indeed how to best measure the effects of computer generated counter-narratives (i.e. textual responses to hate messages), in order to identify the most promising approaches. This thesis will explore this topic across NLP, NLG and complex networks in order to combine content-based, emotion-based and network-based metrics and apply them effectively to fight online hate via analysis of Social Media content spreading.

Machine and deep learning methods are entering also the industrial sector to automatise 3D monitoring and analysis tasks. The research should investigate the use of AI-based methods to boost photogrammetric 3D inspections for industrial quality control operations. Innovative and advanced AI-based solutions should be developed in order to inspect non-collaborative surfaces (reflective, transparent, etc.) and derive precise 3D results useful for quality control.

Machine learning and deep neural networks are extensively and successfully used to process multimodal data (e.g., audio, video, environmental data) on powerful computers. At the same time, several challenges still need to be solved to bring AI on low consumption devices (e.g., end nodes in an IoT) with limited resources. Recently TinyML approaches are emerging to distribute the intelligence at the far edge in the edge-to-cloud continuum. Exciting research scenarios emerge, spanning from novel, innovative hardware for always-on and event-based sensing to tiny deep learning solutions for inference on resource-constrained platforms based on distillation, quantization, or neural architecture search. The complexity grows if we want to move learning to the edge. Motivated by these scenarios, the research aims to (i) define novel hardware/software approaches to optimize AI at the very edge on energy-efficient embedded devices, in particular for audio processing and/or computer vision; (ii) to explore the potential of distributing and fuse the intelligence in heterogeneous nodes of an IoT (iii) to demonstrate the advantages of the investigated approaches in real-world application scenarios, such as those of smart cities.

Data is often collected at the edges of the network but processed centrally fueled by the availability of computing power provided by the cloud. However, the edge of wireless networks can play a role as a distributed platform for ML mitigating the latency and privacy concerns as well as alleviating backhaul network from the transmission of data to the cloud.

The main goal of this PhD is to investigate the impact of bringing learning at the edges of wireless networks, considering an edge-cloud network which is AI aware and where machine learning algorithms interact with the physical limitations of the wireless medium.

The ability of modeling, understanding and predicting human behaviors, mobility routines and social interactions is fundamental for computational social science and has a range of relevant applications for individuals, companies, and societies at large. In this project, the goal is merging approaches from machine learning and network science (e.g., graph neural networks, multi-agent deep reinforcement learning, etc.) and using data on mobility routines (e.g., GPS and other mobile phone data), face-to-face interactions and communication data in order to develop methods for quantify daily habits, individual dispositions and traits, and behavioral changes. A special attention will be given to the changes on daily human behaviors due to the emergence and spread of the Covid-19 pandemic and other shocks. The Ph.D. project will be conducted within the FBK MobS research unit but with collaborations with several international groups (i.e., MIT Connection Science) and with the ELLIS program of the Human-Centric Machine Learning.

The European open data policies have led to the definition of the concept of data space: ecosystem of data within a specific application domain and based on shared policies and rules where users are enabled to access data in a safe, transparent, reliable way, easy and unified.
In this project, the goal is to provide Human-Centered AI tools capable of enabling a data space for mobility , in the context of the European green deal, keeping a balance between users' freedom and companies' constraints.

In the context of the green deal transition and climate change we are looking for candidates willing to develop novel methodologies based on machine learning, deep learning, pattern recognition and artificial intelligence for information extraction, classification, target detection and change detection in long and dense timeseries of remote sensing images.

The candidate will be requested to deal with multi-/hyper-spectral images acquired by passive satellite sensors and/or Synthetic Aperture Radar (SAR) images acquired from active systems for Earth Observation. Among the others, data from ESA Copernicus (Sentinels), ASI PRISMA and COSMO-SkyMed will be considered. The goal is to design novel methods able to use temporal correlation to model landcover behaviors, changes and trends for a better understanding of phenomena over the past and the future for detecting trends and changes for modeling and understanding their impacts on climate and environment.
Besides the requirements established by the rules of the ICT school, preferential characteristics for candidates for this scholarship are:

• master degree in Electrical Engineering, Communication Engineering, Computer Science, Mathematics or equivalents;

• knowledge in pattern recognition, deep learning, image/signal processing, statistic/remote sensing, passive/active sensors.

The need to translate audio input from one language into text in a target language has dramatically increased in the last few years with the growth of audiovisual content freely available on the Web. Current speech translation (ST) systems are now required to be flexible and robust enough to operate in different application scenarios. On one side, the industry calls for key features like real-time processing, domain adaptability, extended language coverage, and the capability to adhere to application-specific constraints. On the other side, society calls for new efforts towards inclusiveness with respect to specific categories and groups (e.g. gender-sensitivity, customization to the needs of impaired users). Both industry and society face the orthogonal challenges posed by the variability of audio conditions (e.g. background noise, strong speakers’ accent, overlapping speakers). The objective of this Ph.D. is to make ST flexible and robust to these and other factors.

Conversational agents are designed to interact with users through various communication channels, such as social media platforms, using natural language. Recently neural end-to-end systems have started to be tested to fight misinformation using argument generation to debunk fake news. Still, Neural Language models suffer from limitations such as hallucination and knowledge lack. Scaling to credible, up-to-date and grounded arguments requires world and domain knowledge together with a deep understanding of argumentative tactics. The goal of this PhD Thesis is to overcome the shortcomings of traditional neural language models, by incorporating several knowledge sources, argumentation and domain features into a constrained generation pipeline.

The research project will focus on the development of tiny machine learning models for learning continuously over time and under domain shift.  The research will focus on developing compact deep learning models (i.e. with reduced memory footprint and computational cost) for domain adaptation and continual learning. Techniques for network pruning and Neural Architecture Search methods will be investigated.

The goal of the thesis is to develop AI models for speech processing which are aware of the computational resources and of the application requirements and are capable of dynamically adapting in order to meet such limitations. This entails not only the search for a trade-off between resources and inference performance but also the possibility to dynamically exploit additional computational resources, eventually expanding the model. The project will address both training and inference phases, starting from state of the art supervised techniques as model compression, neural architecture search, distillation and continual learning and pushing them towards continuous and unsupervised solutions.

Climate change and its impact on countless sectors of society has enormously increased the demand for a comprehensive, robust, timely and reliable climate data analysis that provides support to the adaptation and mitigation policies. Earth System Models (ESMs) that faithfully simulate the cycle of the different components of the Earth System (atmosphere, hydrosphere, cryosphere, biosphere) are the key to address the complex challenges the society is facing, and their development requires expertise at the border between physics and computer science. During the PhD the student will be guided in exploring and applying Artificial Intelligence methods for the parametrization of physical processes, leveraging explainable AI and physics-informed machine learning and HPC-enabled large scale data understanding and processing, which try to blend machine learning with physical knowledge to achieve solutions that are physically more consistent.
The activity will be carried out in collaboration with Fondazione Bruno Kessler and within the activities of Earth & Climate Spoke of the National Center for High-Performance Computing (HPC). Candidates familiar with physical process simulations are welcome, and basic knowledge of Machine Learning/Deep Learning is recommended.

The need to translate audio input from one language into text in a target language has dramatically increased in the last few years with the growth of audiovisual content freely available on the Web. Current speech translation (ST) systems are now required to be flexible and robust enough to operate in different application scenarios. On one side, the industry calls for key features like real-time processing, domain adaptability, extended language coverage, and the capability to meet application-specific constraints. On the other side, society calls for new efforts towards inclusiveness with respect to specific categories and groups (e.g. gender-sensitivity, customization to the needs of impaired users). Both industry and society face the orthogonal challenges posed by the variability of audio conditions (e.g. background noise, strong speakers’ accent, overlapping speakers). The objective of this PhD is to make ST flexible and robust to these and other factors.

The goal of the thesis is to develop AI models for speech processing which are aware of the computational resources and of the application requirements and are capable of dynamically adapting in order to meet such limitations. This entails not only the search for a trade-off between resources and inference performance but also the possibility to dynamically exploit additional computational resources, eventually expanding the model. The project will address both training and inference phases, starting from state of the art supervised techniques as model compression, neural architecture search, distillation and continual learning and pushing them towards continuous and unsupervised solutions.

Machine and deep learning methods are entering also the industrial sector to automatise 3D monitoring and analysis tasks. The research should investigate the use of AI-based methods to boost photogrammetric 3D inspections for industrial quality control operations. Innovative and advanced AI-based solutions should be developed in order to inspect non-collaborative surfaces (reflective, transparent, etc.) and derive precise 3D results useful for quality control.

Automated planning is successfully used in some application areas for the synthesis of plans to control complex systems. Despite significant progress in the literature, scalability is still a major problem that hinders adoption of planning in a wider range of domains.
In this PhD research in the area of integrative AI, the candidate will study methods for combining modern deep learning approaches with symbolic AI for the automatic adaptation of planning tools. In particular, he/she will develop algorithms to automatically specialize planners on specific domains to improve on scalability by exploiting the characteristics of the target domain extracted automatically by means of machine learning.

Machine learning and deep neural networks have been extensively and successfully used to process multimodal data (e.g., audio, video, environmental data) on powerful computers. At the same time, several challenges remain open to move AI onto low-consuming, resource-constrained devices (e.g., end nodes in an IoT). Recently TinyML approaches are emerging to distribute the intelligence to the far edge of the edge-to-cloud continuum. Exciting research scenarios emerge, spanning from tiny deep learning solutions for inference on resource-constrained platforms based on distillation, quantization, or neural architecture search going up to combing software techniques with novel, innovative hardware supporting TinyML. The complexity grows if we consider moving learning to the edge in order to benefit from the opportunities offered by connected, distributed devices. Motivated by these scenarios, the research aims (i) to define novel hardware/software approaches to optimize AI at the very edge on energy-efficient embedded devices, in particular for audio processing and/or computer vision, but not only; (ii) to explore the potential of distributing and fusing intelligence from heterogeneous nodes of an IoT (iii) to demonstrate the advantages of the investigated approaches in real-world application scenarios, such as those of smart cities.

In the context of the green deal transition and climate change we are looking for candidates willing to develop novel methodologies based on machine learning, deep learning, pattern recognition and artificial intelligence for information extraction, classification, target detection and change detection in long and dense timeseries of remote sensing images.
The candidate will be requested to deal with multi-/hyper-spectral images acquired by passive satellite sensors and/or Synthetic Aperture Radar (SAR) images acquired from active systems for Earth Observation. Among the others, data from ESA Copernicus (Sentinels), ASI PRISMA and COSMO-SkyMed will be considered. The goal is to design novel methods able to use temporal correlation to model landcover behaviors, changes and trends for a better understanding of phenomena over the past and the future for detecting trends and changes for modeling and understanding their impacts on climate and environment.
Besides the requirements established by the rules of the ICT school, preferential characteristics for candidates for this scholarship are:
• master degree in Electrical Engineering, Communication Engineering, Computer Science, Mathematics or equivalents;
• knowledge in pattern recognition, deep learning, image/signal processing, statistic/remote sensing, passive/active sensors.

In the last decade, Social Media Platforms have become our main communication hub, encompassing both our personal and our public life. As online social networks have established themselves as important sources or information, they have rapidly changed the landscape of the news media ecosystem on a global scale. News is no longer exclusively broadcast by established sources. Within the participatory environment of these platforms, new opinion leader often actively creates and disseminate news without the restrictions posed upon classical media channels, and often reach large audiences. In this PhD project, we want to investigate how the structural and functional characteristics of online communication networks influence the circulation of news with a focus on disinformation and more broadly junk news. We further want to investigate how such unreliable information diffuses differently across heterogeneously formed communities and characterize the behavior surrounding their reception. We will use methodologies coming from network science, data science and complexity science, integrating the insight about the role taken users that can be obtained by the analysis of the communication network and the associated spreading dynamics with insight about the stance of those users that be automatically extracted using NLP methods.

Supervised learning is a popular mechanism to teach machines vision-based tasks and skills. However, human supervision is a bottleneck for building generic machines that can operate across different contexts, environments and applications. Ideally, machines should develop their own effective and possibly creative strategies for using the sensed data and their experience to continually learn without humans at their side. The research activities related to this PhD position will focus on building novel deep learning-based vision algorithms to teach machines to seamlessly understand environments through 2D or 3D perception.

Precise monitoring of the highly dynamic space radiation environment around Earth is crucial for spacecraft safety.
It supports development of solar particle flux models and allows studies of space weather and of the interaction of radiation belts with Earth's lithosphere.
The project activities include the study of a flat detection geometry to reduce the size of the low energy particle detector.
The project also includes the parametric characterization of the sensors, the development as well as  testing with particle beams of a detector prototype and more in general the integration of the payload.

The LHC experiments produce about 90 petabytes of data per year. Inferring the nature of particles produced in high-energy collisions is crucial for both probing the Standard Model with greater precision and searching for phenomena beyond the Standard Model. In this context, event selection is becoming more difficult than ever before and requires expertise at the border between physics and computer science. During the PhD the student will be guided in exploring and designing Deep Learning algorithms to tackle this problem learning how to apply rigorous Data Science methodologies. The activity will be carried out in collaboration with INFN-TIFPA, Fondazione Bruno Kessler and within the ATLAS experiment at the LHC. Candidates familiar with High Energy Physics are welcome, and basic knowledge of Machine Learning/Deep Learning is recommended.

Research activity conducted during the Ph.D. will focus on the development of mathematical and statistical models to investigate the transmission of emerging and re-emerging pathogens in human populations. This may include the analysis of spatio-temporal patterns characterizing an observed epidemic, the estimation of the contribution of different settings (e.g., households, schools, workplaces, hospitals) in the spread of an infectious diseases, the forecast of potential epidemic trajectories, and the exploration of alternative intervention scenarios (e.g., social-distancing measures, vaccination). Envisioned approaches range from the development and simulation of mechanistic transmission models to the use of statistical inference applied to epidemiological data.

In the last decade, Social Media Platforms have become our main communication hub, encompassing both our personal and our public life. As online social networks have established themselves as important sources or information, they have rapidly changed the landscape of the news media ecosystem on a global scale. News is no longer exclusively broadcast by established sources. Within the participatory environment of these platforms, new opinion leader often actively creates and disseminate news without the restrictions posed upon classical media channels, and often reach large audiences. In this PhD project, we want to investigate how the structural and functional characteristics of online communication networks influence the circulation of news with a focus on disinformation and more broadly junk news. We further want to investigate how such unreliable information diffuses differently across heterogeneously formed communities and characterize the behavior surrounding their reception. We will use methodologies coming from network science, data science and complexity science, integrating the insight about the role taken users that can be obtained by the analysis of the communication network and the associated spreading dynamics with insight about the stance of those users that be automatically extracted using NLP methods.

Deep supervised learning methods, combined with clinical data, have been very successful in predicting disease progression, estimating risk factors and other outcomes of clinical interest in different areas of medicine. However, predictions alone, while useful are not sufficient. Methods that can recommend whether the patients should be treated, and in what way, are necessary. This is known as treatment effect estimation and requires understanding causal relationships between variables found in observational data, such as electronic health records.

In this respect, the work will be focused on developing deep learning methods that can learn causal representations from clinical data, and then provide causal reasoning, including prediction of counterfactuals. The work will build on top of existing approaches, such as Structural Causal Models, as well as more recent approaches, including Recurrent Marginal Structural Networks and Counterfactual Recurrent Networks.

The candidate will have a strong background in machine learning, statistics, mathematics or a related field and will work with real-world patient data related to chronic diseases and critical care. During the PhD the candidate will have the opportunity to collaborate with some of the leading clinical experts in the USA (including from MIT, Mayo Clinic, Cleveland Clinic) as well as experts from leading European institutions.

Energy communities are at the forefront of the EU Green Deal strategy. Since 2016 a number of works have been done by FBK-SE covering the planning of several municipalities and regions based on EnergyPLAN+MOEA. However, these case studies are lagging behind in respect to some aspects: (I) full integration of the multiple decision variables that maximize flexibility, (II) Multi-Node solutions that enhance the synergies between different local, regional, national and transnational scales, (III) holistic approaches among energy-environment-economy-society, (IV) interaction with geospatial models dedicated to land-use, urban planning, mobility, etc., (V) embedding of both climate mitigation and adaptation. Moreover, the current approach is missing the integration of a power flow analysis. Therefore, the overarching goal of this PhD is to develop an innovative multi-objective optimization framework for Smart Energy Communities, including energy and power flow modeling.

The aim of this PhD project is to investigate new technology-based approaches to support adaptive educational paths and/or (partially-)automated personalized support for teachers and/or students. These may include end-user programming for teachers and students for personalizing and co-creating learning activities; human-centred approaches to the design of AI-systems for education; initiatives for support young people and educators that do not have specialized knowledge or technical skills of AI to personalize and tailor AI-based systems to their needs.

The ideal candidate has a background in Computer Science, Psychology or Cognitive Science. Knowledge and experience in data science and basics of Artificial Intelligence are required as well as competences with educational theories. Experience with design of interactive digital technologies, conduction of experimental and in-the-wild studies, and international mobility are a plus for the application and should be acquired during the Phd training

The aim of this PhD project is to investigate new technology-based approaches to support adaptive educational paths and/or (partially-)automated personalized support for teachers and/or students. These may include end-user programming for teachers and students for personalizing and co-creating learning activities; human-centred approaches to the design of AI-systems for education; initiatives for support young people and educators that do not have specialized knowledge or technical skills of AI to personalize and tailor AI-based systems to their needs.

The ideal candidate has a background in Computer Science, Psychology or Cognitive Science. Knowledge and experience in data science and basics of Artificial Intelligence are required as well as competences with educational theories. Experience with design of interactive digital technologies, conduction of experimental and in-the-wild studies, and international mobility are a plus for the application and should be acquired during the Phd training

The aim of this PhD project is to investigate new technology-based approaches to support adaptive educational paths and/or (partially-)automated personalized support for teachers and/or students. These may include end-user programming for teachers and students for personalizing and co-creating learning activities; human-centred approaches to the design of AI-systems for education; initiatives for support young people and educators that do not have specialized knowledge or technical skills of AI to personalize and tailor AI-based systems to their needs.

The ideal candidate has a background in Computer Science, Psychology or Cognitive Science. Knowledge and experience in data science and basics of Artificial Intelligence are required as well as competences with educational theories. Experience with design of interactive digital technologies, conduction of experimental and in-the-wild studies, and international mobility are a plus for the application and should be acquired during the Phd training

The Limadou project gathers some Italian institutions participating in the China Seismo Electromagnetic Satellite (CSES) mission. CSES consists of a constellation of satellites, designed to pursue the deepest campaign of observation of the ionosphere. One of the most important scientific goals of the mission is to look for correlations between transient phenomena in the ionosphere and seismic events. Among payloads, a set of particle detectors is devoted to the detection of charged particles trapped in the Van Allen Belts, to monitor the solar activity and to measure galactic cosmic rays of very low energy. The APP group of the Physics Department in Trento looks for candidates to a PhD programme on the analysis of the scientific data from the payloads on board the CSES-01 and those to be launched on board the satellite CSES-02 in 2022. The student will focus on time-series analyses and participate in the development of the event reconstruction software. These studies will be carried out using the most modern machine learning techniques for clustering and anomaly detection, using full information from CSES payloads. The activity will be carried out in collaboration with INFN-TIFPA, Fondazione Bruno Kessler and the Institute of the High Energy Physics of Beijing. Candidates familiar with the experimental techniques for the detection of charged particles in space are welcome, as well as  basic knowledge of Machine Learning/Deep Learning is recommended.

Water phase changes (evaporation, condensation, freezing) are ubiquitous phenomena of great importance for living beings and in engineering applications. The structure (micro and nano) and chemistry of surfaces control the kinetics and dynamics of these transitions. Plants, for example, offer numerous examples of self-cleaning, antifreeze and water-collecting 1 properties developed over millions of years of evolution. Engineered anti-frosting surfaces find applications in aerospace (ice accretion on aircrafts), heat exchangers (refrigerators), wind turbines and power lines. Structured surfaces that increase evaporation and condensation efficiency are a challenge for Loop Heat Pipes (LHP) and Vapour Chambers that cool electronics on space stations (in microgravity conditions) or in the electronic devices we use on a daily basis. Surfaces that can efficiently collect dew and fog provide a source of water in arid environments and can improve the water recovery system of space stations.

This project will extend the studies carried out during the previous PhD scholarship (within cycle 34, in collaboration with FBK) which focused on anti-frosting and water-harvesting surfaces. The research activity will concern the theoretical study, fabrication, characterisation and experimentation of micro- and nanostructured surfaces with applications in aerospace and energy efficiency. In particular, phenomena of spontaneous jumps of condensation droplets on hydrophobic surfaces, distant coalescence on hydrophilic surfaces and freezing of droplets will be studied. Fabrication techniques may range from micro- and nanolithography, focused ion beam, etching, chemical deposition processes, and polymer moulding. The expected outputs are patents and publications on high impact journals in the field.

Low Gain Avalanche Diodes (LGADs) are silicon sensors that feature internal charge gain. These sensors were initially developed to provide the time information of tracks at high luminosity colliders, with performances reaching single hit time resolutions of a few tens of picoseconds for minimum ionizing particles. These timing capabilities can find applications in spaceborne experiments like: particle identification through time of flight, distinction between incoming and outgoing particles, identification of splash back and punch through of showers in the calorimeter systems, identification of electromagnetic and hadronic showers by observation of the splash back and punch through from calorimeters. A first production of LGADs dedicated to space applications was completed at Fondazione Bruno Kessler (FBK) and is currently being characterized. The activities of this position will be focused on completing the characterization of this first batch, and on the qualification tests to determine whether the LGAD sensors are flight-ready. The sensor characterization will be mainly performed in the laboratories of University of Trento and FBK. The lessons learned in the sensor characterization and qualification will be reflected in the design of future sensors dedicated to spaceborne experiments. Within the timeframe of this position, a second batch of LGADs for space will be produced and characterized.

The phd will investigate new techniques for model-based system and software engineering and formal methods to support the design, mission preparation and operations of space systems. The potential research directions include faul detection, isolation, and recovery for satellites; system level diagnosis and diagnosability based on telemetry; digital twins for satellites.

In the last few years, several companies and researchers have designed AI-based approaches for assessing and selecting the best candidates and talents based on allegedly objective performance criteria. Ideally, the usage of machine learning in assessment has the objective of mitigating the biases which often affect the hiring decisions conducted by human recruiters.
However, several studies have shown that also machine learning algorithms can be characterized by biases. Hence, the goal of this PhD project is to develop and evaluate innovative, fair and explainable machine learning models for inferring candidates’ employability, performance, and individual characteristics.
In particular, the candidate is expected to develop innovative natural language processing and/or multimodal (audio-video) algorithms to extract and to infer information from professional resumes and/or video job interviews in order to evaluate skills, possible future performances and individual characteristics of candidates. Moreover, state-of-the-art and innovative approaches to machine learning fairness will be implemented and evaluated by the student. The outcome of the student may consist in research prototypes that will be tested on Gi-Group data as well as on patents and scientific publications in top-tier conferences (AAAI, ACL, IJCAI, AIES, FaccT, etc.) and journals.
The intellectual property of the research results that will derive from the activities carried out by the doctoral student is owned by FBK and the Company.

Sensing devices interact in complex environments and their miniaturization and portability call for small form factor and low power consumption. Edge AI is therefore of fundamental importance for robustness, privacy, and long battery operation. The application of AI algorithms to image sensors further constrains technological solutions in resources due to the large amount of generated data. The objective of this project is to study HW-friendly solutions exploiting digital parallel and dedicated HW/SW architectures aimed at solving specific sensing use cases with high efficiency in area and power. The student will have to tackle the challenges with a multidisciplinary point of view, from the study of the sensing problem to the modelling of the solution, and a special focus on the integrated circuit design aspect. Technically speaking, the work will include the development of hardware-friendly algorithms based on deep learning, high-level modeling of an imaging system, simulations of the system behavior in real case scenarios, implementation of the system in a mixed hardware/software platform (which might include PC, SoC, FPGAs, integrated circuits) and characterization of developed system in the lab and on the field. The student will interact with experts in the fields of computer vision and CMOS image sensors gaining a unique combination of background knowledge. The expected outcome is a highly optimized architecture for an imaging system that combines software, firmware and hardware solutions to solve specific, market-driven needs in the consumer electronics, automotive, smart city, robotics, digital industry fields. The intellectual property of the research results that will derive from the activities carried out by the doctoral student is owned by FBK and the Company.

THz/IR sensing and imaging systems can expand the capabilities of portable devices beyond the human vision. The objective of this project is to study sensing solutions operating from the THz to the infrared region exploiting antenna-coupled field-effect transistors, including the study and optimization of the devices and the electronic design of integrated readout and control. The student will have to tackle the challenges with a multidisciplinary point of view, from the quasioptical and electromagnetic point of view, the device behaviour in terms of response to an incoming signal and in terms of noise, and a special focus on the integrated circuit design aspect.
Technically speaking, the work will include high-level modeling of devices and sensor architectures, with emphasis on sensitivity and propagation of noise through the readout chain, development of novel pixel and readout architectures, circuit design including schematic and layout, and characterization of fabricated devices in the lab.
The student will interact with experts in the fields of image sensors, THz detectors, and analog circuit design, gaining a unique combination of background knowledge.
The expected outcome is the realization of state-of-the-art image sensors and their validation in a real use-case scenario.
The intellectual property of the research results that will derive from the activities carried out by the doctoral student is owned by FBK and the Company.

In industrial realities, management systems have a categoriesed structure that studies the scope of application discriminating between the different production phases, the type of resources and the dedicated operating departments. The implementation of a quality management system integrates the Quality principle into manufacturing activities is an opportunity to guarantee the quality of research results and to improve and gain recognition for the work done in a research laboratory. In the context of a clean room of R&amp;D production, we want analysing some process flows, which have been identified for this purpose, having as its ultimate goal the CR's already active quality system management model optimisation. This thesis work involves the use of various evidence-based and statistical tools for the definition and visualisation of processes, the identification of possible failures or criticalities and the definition of consequent corrective actions. This approach will define a new model for assessing and managing non-conformities, which are already dealt with the current quality system. The novelty introduced is the development of a management system starting from the knowledge of industrial realities certified and the codification of know-how developed in the MNF CR itself. A practical system declination is the proposal of preventive and corrective actions as tools for non-conformities and criticalities handling as highlighted in the monitoring of process activities. The thesis is organised in two phases. The first phase analyses Dry Etching process as the case study chosen for the definition of the model described above. This choice turns out to be sufficiently complex to represent a self-consistent system that we expect to capture the variability of operational parameters like a general model. For this reason, the second phase involves the study of the generalisation of the validity of the elaborated model to a second domain, relating to the field of deposition processes.

Digital technologies play an ever increasing role in all aspects of human society; this induces a wide range of changes, collectively referred to as Digital Transformation, that, far from being only technological, also cover cultural, organizational, social, managerial aspects of our life.

Artificial Intelligence is a key technology for digital transformation, thanks to its capability to extract information and knowledge from data; this requires the capability to open, analyze and exploit all data available on a given phenomenon, data that are often highly heterogeneous, scattered, and coming from different sources (e.g. open, sensor, free, closed, linked data). This thesis will concentrate on developing a data-driven computational framework, based on AI approaches, able to perform data analysis and prediction in the setting just described. The framework will be developed in the scope of the Digital Hub, a digital platform jointly developed by Dedagroup and Fondazione Bruno Kessler to address digital transformation in different application domains, including Public Administration, Digital Finance, Digital Industry. The validation of the framework will be performed addressing problems in these application domains, by exploiting the data sets and services integrated in the Digital Hub

This PhD aims at investigating and training neural architectures to build large multimodal sequence-to-sequence models able to encode input speech and text in a common space, and also able to decode output text or speech from a common representation. Such a multimodal architecture, if then made multilingual, could become the future unified foundation model for building automatic speech recognition (ASR), text-to-speech synthesis (TTS), speech-to-text translation (S2T) and speech-to-speech translation (S2S) systems from a single backbone.

Downscaling and upscaling of fields of atmospheric variables from modelling and observations by means of Artificial Intelligence techniques. Climate changes and their effects through weather modifications have an enormous impact on countless sectors of society. As a cosequence wether services are facing an increasing demand for comprehensive, robust, timely, reliable and high-resolution information from either moinitoring systems, or weather forecasts or climate projections that provide support to the adaptation and mitigation policies. High-resolution (in space and time) fields are a key tool towards addressing the complex challenges society is facing. Their development requires an intedisciplinary expertise between meteorology, physics, applied mathematics and computer science. The candidate will develop and apply new concepts in the application of Artificial Intelligence (Machine Learning and Deep Learning) for the spatial and temporal downscaling of forecasts, observations and climate projections from coarse-grained sources and vice versa. The activity will be carried out in collaboration with Fondazione Bruno Kessler and within the activities of Earth & Climate Spoke of the National Center for High-Performance Computing (HPC).

Model-based safety assessment (MBSA) is a growing research area in the design of complex safety-critical systems. Starting from requirements and formal models of the system under analysis, automated techniques and tools are used to analyze system correctness and dependability, and to support its certification, automatically constructing safety artifacts such as Fault Trees and FMEA tables.

Objective of the study is to lift MBSA techniques from finite-state systems to the case of hybrid systems that include continuous time and complex dynamics.  The study will investigate three related directions. First, model extension, i.e., the generation of models encompassing faulty behaviors from nominal models, based on a library of predefined faults, specifying the effects and dynamics of faults. Second, the design of engines for the verification and synthesis of safety-related artifacts, based on state-of-the-art parameter synthesis techniques. Finally, the use of contract-based analysis techniques, which exploit the system architecture to perform safety assessment hierarchically.

The Study will be conducted as part of several ongoing research projects carried out at FBK, such as VALU3S (EU funded) and COMPASTA (funded by the European Space Agency).

Cyber deception is a defense strategy, complementary to conventional approaches, used to enhance the security posture of a system. The basic idea of this technique is to deliberately conceal and/or falsify a part of such system by deploying and managing decoys (e.g. "honeypots", "honeynets", etc.), i.e., applications, data, network elements and protocols that appear to malicious actors as a legitimate part of the system, and to which their attacks are misdirected.  The advantage of an effective cyber deception strategy is twofold: on one hand, it depletes attackers' resources while allowing system security tools to take necessary countermeasures; on the other hand, it provides valuable insights on attackers' tactics and techniques, which can be used to improve system's resilience to future attacks and upgrade security policies accordingly.

Although cyber deception has been successfully applied in some scenarios, existing deception approaches lack the flexibility to be seamlessly operated in highly distributed and resource-constrained environments. Indeed, if virtualisation and cloud-native design approaches paved the way for ubiquitous deployment of applications, they widened the attack surface that malicious actors might exploit. In such a scenario, it is practically unfeasible to try to deploy decoys for each and every system's service or application without dramatically depleting resources, especially in edge scenarios, where these are scarcely available.

This calls for a novel approach to cyber deception combining security, networking, cloud and AI technologies, that takes the tradeoff between security and efficiency into account and makes deception strategies more effective in cloud-to-edge environments. The PhD project will tackle the above mentioned challenges from different perspectives, including the dynamic and automated orchestration of decoys, the design and implementation of lightweight and flexible honeypots, the proposition and evaluation of relevant performance indicators and the integration and interaction with DevOps and SecOps tools.

Machine Learning (ML) is nowadays a consolidated technology embedded in various domains of computer science and information technology. In the recent past years, ML has revolutionised cybersecurity applications, with excellent results in various application areas such as: encrypted traffic classification, intrusion detection and prevention, anomaly detection in industrial control systems, identification of malicious software (or malware), among others.

One important research subfield of ML is called Explainable Machine Learning, which relates to understanding the ML model behaviour by means of various techniques such as feature importance scores, influential training data, etc,. Given the complexity of some black-box ML models, it is inherently difficult to understand why they behave the way they do. Understanding how a ML model works and how it takes its decisions is paramount in network security. Indeed, the ability to understand why an event is classified as benign or malicious by an ML-based intrusion detection system allows the ML practitioner to take the necessary counteractions to reduce false positive and false negatives rates, and to make the system more robust to Adversarial Machine Learning attacks.

The objective of this PhD project is to perform fundamental research in the field of ML explainability (understanding how ML algorithms reason their outputs) and to propose novel tools and methodologies for ensuring good performance of ML-based security systems under various working conditions.

The growing demand for low-cost sensors is gaining momentum in various application fields. One of the main drawbacks of low-cost gas sensors, such as chemoresistive ones, is the lack of selectivity, which limits the use of these devices for monitoring specific gases in a complex environment. To overcome this drawback, research currently focuses mainly on the development of novel sensing materials, where the introduction of specific functionalization could lead to increased device selectivity. On the other hand, another interesting approach that can be exploited is the introduction, in the device packaging, of membranes with selective permeation properties, leading to an improvement of the device performance. This approach has the great advantage of not intrinsically modifying the gas sensor, whose development process is well-established, and of being able to tailor the development of the selective membrane according to the gas to be analyzed. In particular, the PhD project proposed here focuses on the development of membranes for the selective permeation of H2, an increasingly energy-important fuel gas for which there is currently a lack of reliable and low-cost technologies for its monitoring.

The affective dimension of word meaning often forms part of our reservoir of common-sense knowledge, and it is reflected in the way we use words. This project aims at producing and evaluating new technologies for recognition of emotional language and possibly other subtle pragmatic aspects of communication. Because there are diverse subtilties in emotional expressions in different languages, the project will devote particular attention in approaching the problem from a multilingual point of view.

Required skills:
Good familiarity and expertise with Computational Linguistics techniques. Experience in machine learning. Good programming skills.

Human-human dialogues are characterized by collaborative behaviours, through which interlocutors achieve their communicative goals. As an example, proactivity (i.e., anticipating user needs during dialogue) and grounding (e.g., posing clarification questions) are two relevant cases that have been investigated from a linguistics perspective. However, such collaborative behaviours are still largely absent in current neural dialogue models. There are several open research challenges in this direction, including investigating how dialogue systems can learn when and how to be collaborative, depending on the dialogue context, and how do we evaluate whether collaborative behaviours have improved the efficacy of dialogue. This PhD project addresses collaborative behaviours in conversational agents from a computational perspective, exploiting the integration of machine learning approaches based on neural models, reinforcement learning, and knowledge-based techniques.

One of the pillars of healthcare digital transformation focuses on the integration of AI-based solutions within the clinician-patient relationships with the aim of monitoring and/or supporting them towards the achievement of healthy functional status.

Examples of these systems are: (I) virtual coaches to support remote monitoring and recommendations for patients affected by nutritional chronic diseases or to support the prevention of the onset of such diseases; (ii) telehealth solutions to enhance care capabilities of health organizations; and, (iii) tools to orchestrate care pathways involving, beside patients, multiple clinical actors.

This Ph.D. works within this context with the aim of designing novel AI-based approaches to trigger the implementation of the next-generation virtual digital assistants.

The area of intervention is very broad since the research areas involved are, for example, knowledge management, human-computer interaction, pervasive computing, machine learning, probabilistic graphical model, natural language processing, and planning.

For this reason the Ph.D. candidate will have the opportunity to explore the virtual digital assistants domain in order to analyse current open challenges, to decide which ones to address and which AI-based approaches she/he will use to tackle such challenges.

The goals of the PhD research are: (i) to study, develop and validate innovative solutions based on AI algorithms to extract geometric and semantic information from digital data (images and 3D models) of cultural heritage; (ii) to propose and test alternative methods that allow to apply machine / deep learning algorithms in contexts with little data availability and with noisy classes, by exploiting integrative AI methods; (iii) to analyze, realize and demonstrate new methods to improve the transparency, interpretability and explainability of AI methods applied to Cultural Heritage 3D data.

The research should tackle the problems with a holistic and integrative approach, considering multi-GPU approaches, increasing learning capabilities and allowing to handle data with noise. Predictive solutions will serve to better analyze, preserve and enhance the Cultural Heritage, as well as to develop VR / AR solutions to support the tourism sector.

Learning-based algorithms for 3D object description, recognition and retrieval suffer from lack of annotated data, unbalanced classes, computationally inefficient processing pipelines and poor generalization ability across different application domains. All these factors together often hamper the employment of 3D processing pipelines in large-scale real-world applications in urban and environmental contexts.

The goal of this Ph.D. position is to conduct research on novel and efficient algorithms for 3D data semantic segmentation and classification using integrative AI approaches that can effectively replace traditional hand-crafted modules to ultimately improve performance, ease deployment and foster scalability. The research should integrate traditional 3D classification methods (RF, 3DCNN, MLP, etc.) with symbolic approaches (KBANN, LTN, etc.) in order to enhance learning capabilities, handle noisy and multi-modal data and deliver a hybrid method able to constraint predictions with a-priori knowledge expressed in terms of logical formulas. A research task should also be dedicated to investigate advanced solutions to handle unbalanced classes in 3D classification problems, considering e.g. oversampling and under-sampling techniques, uneven weight distribution, complex loss functions, etc.

Many application domains require the ability to automatically generate a suitable course of actions that will achieve the desired objectives. Notable examples include the control of truck fleets for logistic problems, the organization of activities of automated production sites, or the synthesis of the missions carried out by unmanned, autonomous robots. Planning and scheduling (P&S) are fundamental research topics in Artificial Intelligence, and increasing interest is being devoted to the problem of dealing with timing and resources. In fact, plans and schedules need to satisfy complex constraints in terms of timing and resource consumption, and must be  optimal or quasi-optimal with respect to given cost functions. The Ph.D. activity will concentrate on the definition of an expressive, formal framework for planning with durative actions and continuous resource consumption, and on devising efficient algorithms for resource-optimal planning. The activity will explore the application of formal methods such as model checking for infinite-state transition systems, and Satisfiability and Optimization Modulo Theories, and will focus on practical problems emerging from the flexible manufacturing domain.

The advent of Industry 4.0 has made it possible to collect huge quantities of data on the operation of complex systems and components, such as production plants, power stations, engines and bearings.  Based on such information, deep learning techniques can be applied to assess the state of the equipment under observation, to detect if anomalous conditions have arised, and to predict the remaining useful lifetime, so that suitable maintenance actions can be planned. Unfortunately, data driven approaches often require very expensive training sessions, and may have problems in learning very rare conditions such as faults. Interestingly, the systems under inspection often come with substantial background knowledge on the structure of the design, the operation conditions, and the typical malfunctions. The goal of this PhD thesis is to empower machine learning algorithms to exploit such background knowledge, thus achieving higher levels of accuracy with less training data.

Runtime verification is a light weight verification technique based on the analysis of system logs. A key factor is that the internal state of the system is not observable, but partial knowledge on its behaviour may be available. The thesis will investigate the use of temporal epistemic logics (i.e. logics of knowledge and believe over time) to specify and verify hyperproperties for runtime verification. Different logical aspects, like distributed knowledge and common knowledge, and the communication between reasoning agents, will be used to model hierarchical architectures for fault detection and identification, and for prognosis. Techniques for planning in belief space will be used for the design of fault reconfiguration policies.

Many artifacts in the development process (requirements, specifications, code) tend to become legacy, hard to understand and to modify. This results in lack of reuse and additional development costs. A reverse engineering activity is necessary to understand what the system is doing. Goal of the thesis is to provide automated techniques to analyse the inherent behavior of legacy artifacts, extract interface specifications, and to support re-engineering activities. The thesis will combine techniques from language learning, applicable to black-box artifacts, and formal techniques for the automated construction of abstractions in the form of extended finite state machines.

Simulation tools are fundamental to predict the evolution of pandemic, and to assess the quality of counter-measures, e.g. the effect of travel restrictions on the spread of the coronavirus. However, they come with two fundamental requirements. The first is the need for a fast reconfiguration of the simulation, in order to be able to describe the mutating scenarios of the pandemics. The second is the ability to produce correct and explainable results, so that they can be trusted and independently validated. The topic of this research is to devise a model-based approach that is able to represent at a high-level the features of a generic pandemic, from which an efficient simulator can be produced. Using formal methods, the results of the simulation are guaranteed to be correct by construction, with proofs that can be properly visualized and independently checked. The activity will be carried out as a collaboration of the Center for Health Emergencies (https://www.fbk.eu/it/health-emergencies/), that played a major role during the ongoing pandemics, and the Center for Digital Industry (https://dicenter.fbk.eu/), a leading centers in  model-based design.

Applications that rely on the most modern sensing devices and technologies and combine complex artificial intelligence tasks are now mainstream. It is sufficient to say, “OK-Google/Alexa/Siri switch on the heating system when the temperature is below 18° C” to appreciate the power of the IoT in combination with an Artificial Intelligence engine. However, the typical approach to enable intelligent applications is cloud-centric, meaning that the intelligence (a home assistant) is hosted in the cloud infrastructure, and the sensor data collected by some IoT devices (a microphone array and a temperature sensor) flow from the cyber-physical-system until reaching a remote endpoint to be processed. Finally, the correct command is transmitted to the IoT actuator (a radiator thermostat). Alternative approaches to this are possible, for instance, by considering a more dynamic and configurable intermediate layer placed between the IoT and the Cloud sides, usually dubbed as the Edge layer.

Generally, a configurable edge layer reduces the required bandwidth and latency and improves users’ privacy. Moreover, if portions of the application intelligence could be hosted in this layer, the IoT device lifetime would be enlarged. However, reconfiguring and deploying an end-to-end processing flow that involves the three aforementioned architectural layers poses major challenges. Select a more efficient detection algorithm from a rich machine learning algorithms library and pushing the “deploy” button of an application dashboard to see the selected algorithm up and running more effectively (according to a given metric) on my smart home devices is still a dream, in most of the cases. Moreover, depending on the hardware capabilities, the application requirements in terms of bandwidth and latency, and the accuracy required for the machine learning task to execute, different end-to-end configurations are possible, all sub-optimal and possibly non-dominated in the Pareto meaning.

The subject of this Ph.D. is to investigate and propose novel optimization and assessment methodologies to efficiently sample such a complex design space in target application sectors such as home, industry, manufacturing, farming, etc. The reference technological environment covers (but is not limited to) embedded device software engineering (micropython, mbed OS, C languages and dialects, etc.), machine learning frameworks deployable on tiny devices (tinyML, TensorFlow lite, etc.), edge-based frameworks (eclipse Kura, edgeX Fundry, etc.) and cloud-based IoT platforms and services with AI support and components (MS Azure, AWS Greengrass, ThingsBoard, etc.).

The advent of Industry 4.0 has made it possible to collect huge quantities of data on the operation of complex systems and components, such as production plants, power stations, engines and bearings.  Based on such information, deep learning techniques can be applied to assess the state of the equipment under observation, to detect if anomalous conditions have arised, and to predict the remaining useful lifetime, so that suitable maintenance actions can be planned. Unfortunately, data driven approaches often require very expensive training sessions, and may have problems in learning very rare conditions such as faults. Interestingly, the systems under inspection often come with substantial background knowledge on the structure of the design, the operation conditions, and the typical malfunctions. The goal of this PhD thesis is to empower machine learning algorithms to exploit such background knowledge, thus achieving higher levels of accuracy with less training data.

Simulation tools are fundamental to predict the evolution of pandemic, and to assess the quality of counter-measures, e.g. the effect of travel restrictions on the spread of the coronavirus. However, they come with two fundamental requirements. The first is the need for a fast reconfiguration of the simulation, in order to be able to describe the mutating scenarios of the pandemics. The second is the ability to produce correct and explainable results, so that they can be trusted and independently validated. The topic of this research is to devise a model-based approach that is able to represent at a high-level the features of a generic pandemic, from which an efficient simulator can be produced. Using formal methods, the results of the simulation are guaranteed to be correct by construction, with proofs that can be properly visualized and independently checked. The activity will be carried out as a collaboration of the Center for Health Emergencies (https://www.fbk.eu/it/health-emergencies/), that played a major role during the ongoing pandemics, and the Center for Digital Industry (https://dicenter.fbk.eu/), a leading centers in  model-based design.

Learning-based algorithms for 3D object description, recognition and retrieval suffer from lack of annotated data, unbalanced classes, computationally inefficient processing pipelines and poor generalization ability across different application domains. All these factors together often hamper the employment of 3D processing pipelines in large-scale real-world applications in urban and environmental contexts.

The goal of this Ph.D. position is to conduct research on novel and efficient algorithms for 3D data semantic segmentation and classification using integrative AI approaches that can effectively replace traditional hand-crafted modules to ultimately improve performance, ease deployment and foster scalability. The research should integrate traditional 3D classification methods (RF, 3DCNN, MLP, etc.) with symbolic approaches (KBANN, LTN, etc.) in order to enhance learning capabilities, handle noisy and multi-modal data and deliver a hybrid method able to constraint predictions with a-priori knowledge expressed in terms of logical formulas. A research task should also be dedicated to investigate advanced solutions to handle unbalanced classes in 3D classification problems, considering e.g. oversampling and under-sampling techniques, uneven weight distribution, complex loss functions, etc.

Many artifacts in the development process (requirements, specifications, code) tend to become legacy, hard to understand and to modify. This results in lack of reuse and additional development costs. A reverse engineering activity is necessary to understand what the system is doing. Goal of the thesis is to provide automated techniques to analyse the inherent behavior of legacy artifacts, extract interface specifications, and to support re-engineering activities. The thesis will combine techniques from language learning, applicable to black-box artifacts, and formal techniques for the automated construction of abstractions in the form of extended finite state machines.

Many application domains require the ability to automatically generate a suitable course of actions that will achieve the desired objectives. Notable examples include the control of truck fleets for logistic problems, the organization of activities of automated production sites, or the synthesis of the missions carried out by unmanned, autonomous robots. Planning and scheduling (P&S) are fundamental research topics in Artificial Intelligence, and increasing interest is being devoted to the problem of dealing with timing and resources. In fact, plans and schedules need to satisfy complex constraints in terms of timing and resource consumption, and must be  optimal or quasi-optimal with respect to given cost functions. The Ph.D. activity will concentrate on the definition of an expressive, formal framework for planning with durative actions and continuous resource consumption, and on devising efficient algorithms for resource-optimal planning. The activity will explore the application of formal methods such as model checking for infinite-state transition systems, and Satisfiability and Optimization Modulo Theories, and will focus on practical problems emerging from the flexible manufacturing domain.

Runtime verification is a light weight verification technique based on the analysis of system logs. A key factor is that the internal state of the system is not observable, but partial knowledge on its behaviour may be available. The thesis will investigate the use of temporal epistemic logics (i.e. logics of knowledge and believe over time) to specify and verify hyperproperties for runtime verification. Different logical aspects, like distributed knowledge and common knowledge, and the communication between reasoning agents, will be used to model hierarchical architectures for fault detection and identification, and for prognosis. Techniques for planning in belief space will be used for the design of fault reconfiguration policies.

Over the last years, industries in all application domains witnessed a fast growth in the adoption of AI techniques based on Machine Learning. The availability of large amounts of data allowed the deployment of solutions like predictive maintenance and various kinds of forecasts.

Traditionally, symbolic and model based techniques are important for various tasks within industries: for example, planning and scheduling are used to automatically synthesize work plans and to control robotic machines; diagnosis is used to identify the source of problems; and verification can certify and debug systems and processes at design time.

In this PhD research, the student will be exposed to both the symbolic and learning perspectives with concrete industrial case studies, researching integrative AI solutions that empower symbolic algorithms and techniques with machine learning as well as learning-based solutions with model-based predictions and analyses.

Cloud adoption is on the rise and promises to offer many advantages by leveraging economies of scale; at the same time, new security and privacy challenges arise.   As an example, consider the protection of data; while in transit and at rest, cryptographic  techniques to guarantee confidentiality and integrity are well-understood and readily available for several different use case scenarios in the cloud. The situation is much less clear for data in use, i.e. during computation, although it is fundamental for trusting cloud service providers without taking for granted their unsupported claims about security assurances especially when sensitive (e.g., healthcare or financial) information is being processed.  To achieve this advanced level of data protection, it is fundamental to design and prove the security of technical enforcement mechanisms of confidentiality and integrity policies in Trusted Execution Environments.  For the usability of these mechanisms, fundamental security services (including key management and attestation) must be developed, their security and risk level formally assessed, and their deployment automated.

The research work to be conducted during the thesis aims to make significant contributions to developing methodologies, automated techniques and tools to assist the development of fundamental services for confidential computing solutions in the cloud with a focus on key management, attestation, and their integration with identity management solutions for both users and machines to establish a root of trust with high assurance.  Applications of interest for the integration of foundational services range from confidential AI, databases, and analytics to confidential ledgers and multiparty collaboration of dataset owners.

References
- Giada Sciarretta, Roberto Carbone, Silvio Ranise, Luca Viganò:
Formal Analysis of Mobile Multi-Factor Authentication with Single Sign-On Login. ACM Trans. Priv. Secur. 23(3): 13:1-13:37 (2020)

- Stefano Berlato, Roberto Carbone, Adam J. Lee, Silvio Ranise:
Exploring Architectures for Cryptographic Access Control Enforcement in the Cloud for Fun and Optimization. AsiaCCS 2020: 208-221

- Edlira Dushku, Md Masoom Rabbani, Mauro Conti, Luigi V. Mancini, Silvio Ranise:
SARA: Secure Asynchronous Remote Attestation for IoT Systems. IEEE Trans. Inf. Forensics Secur. 15: 3123-3136 (2020)

The scholarship aims at exploring synergies between AI techniques (both logic-based and machine learning based) and Process Mining tasks such as discovery, compliance checking, prediction, and recommendation. Special emphasis is placed on the identification of challenges posed by low quality, unbalanced, privacy protected, or scarce/missing data which may affect machine learning based solutions.

Real-time measurements and data collections are still an outstanding challenge, which this thesis program will assist in solving both the surveying and data retrieval and referencing aspects. The goal of the PhD research is develop a photogrammetric methodology to survey and model scenes of any shape, resulting in a scaled semantic 3D point cloud of the surveyed environment. The PhD's goal is to resolve two still open research problems: i) to speed up the digitalization process and ii) to aid data retrieval by using tools that automatically analyze data. In the first step of the research, the candidate will work on automatic real-time image orientation based on V-SLAM techniques. The second step will focus on the integration of machine/deep learning techniques to build a real-time image classification process to help with the automatic referencing of data and info on the 3D digital copy.

Cyber-Physical Systems (CPS) are ubiquitous systems that integrate computation, networking and physical processes. The correctness and dependability of their control software is critical in many high-assurance domains such as space, mobility, and energy. However, their design requires complex component-based hybrid models describing the continuous dynamics of the physical components and the discrete interaction with the control and monitoring components. The objective of this PhD research is that of advancing the state of the art in the formal verification of CPS control design models, integrating various techniques such as SMT-based model checking, contract-based design, abstract interpretation, simulation, test-case generation, and fault injection. The candidate will work on theoretical aspects of the problem as well as its practical applications in relevant case studies, drawn from the domains of aerospace, automotive, railways, and energy.

3D scene understanding using learning-based approaches is becoming largely employed in several application sectors including industry, automotive, surveillance and cultural heritage. Computational algorithms that are traditionally used to process 2D images, such as object detection, tracking and segmentation, are nowadays being successfully extended to process 3D data (e.g. point clouds). However, traditional 3D approaches rarely use the image content directly for their task, but often rely on mid-level representations (e.g. voxels, sparse point clouds) that disregard the rich context provided by images. Moreover, the availability of 3D data is limited because the effort for annotationing it is greater than its 2D counterpart. The goal of this PhD position is to advance the state of the art about 3D scene understanding by focusing on the aspects of learning-based approaches that can potentially leverage different sensor modalities and the lack of human-annotated data.

Cyber-physical systems are ubiquitous systems that integrate computation, networking and physical processes. The correctness and dependability of their control software is critical in many high-assurance domains such as space, mobility, and energy. However, their design requires hybrid models that combine continuous dynamics with discrete states and is typically verified with simulation-based testing. The objective of this PhD research is that of advancing the state of the art in the formal verification of hybrid system models, integrating various techniques such as SMT-based model checking, simulation, test-case generation, and fault injection. The candidate will work on both theoretical aspects of the problem, as well as its practical applications in relevant case studies, drawn from the domains of aerospace, automotive, railways, and energy.

Automated Planning is the problem of synthesizing courses of actions guaranteed to achieve the desired objective, given a formal model of the system being controlled. A class of problems particularly interesting for applications is temporal planning (also called planning and scheduling) where the discrete decisions of "what to do" are coupled with the problem of scheduling (deciding "when to do"). Unfortunately, planning and scheduling techniques suffer from scalability issues and are often unable to cope with the complexity of real-word scenarios, despite the plethora of approaches available in the literature. Recently, efforts such as Deepmind AlphaGO and OpenAI Five hit the headlines, with groundbreaking advancements in the field of reinforcement learning. These techniques are able to automatically learn policies to decide what to do in order to achieve the desired goal. However, they offer no formal guarantee and are not model-based. The research objective of this PhD scholarship is to investigate techniques that combine the formal guarantees offered by automated planning and scheduling with the performance and self-improving capabilities offered by recent advances in deep reinforcement learning to construct self-adaptive planners that can improve over time their performance on specific application scenarios.

The increasing complexity of software systems calls for the development of new methods and tools to design and test software systems characterized by high variability from the point of view of the space of the possible functional configurations, the space of the release architectures, and of the aspects related to dynamic reconfiguration. The goal of this PhD thesis is that of exploring new approaches to the testing, verification and validation of these systems that involve the joint use of model-based and AI based techniques such as planning, machine learning and optimization.

Multi-Access Edge Computing leverages the network's edge to store and process data and applications locally, and provide fast reactions and efficient use of network and computing resources. Future communication systems and networks, both terrestrial and non-terrestrial, will increasingly require solutions based on AI/ML, virtualization and softwarisation techniques besides traditional communication technologies. The goal of this PhD Thesis is to design and explore novel approaches that leverage on the close cooperation with these domains from an overall system perspective.

In the context of the green deal transition we are looking for candidates willing to develop novel methodologies based on machine learning, deep learning pattern recognition and artificial intelligence for information extraction, classification, target detection and change detection in long and dense timeseries of remote sensing images.
The candidate will be requested to deal with both multi-/hyper-spectral images acquired by passive satellite sensors and Synthetic Aperture Radar (SAR) images acquired from active systems for Earth Observation. Copernicus data acquired by the new ESA Sentinels will be considered. The goal is to design novel methods able to use temporal correlation to model landcover behaviors, changes and trends for a better understanding of phenomena over the past and the future for a better modeling and understanding of climate change.
Besides the requirements established by the rules of the ICT school, preferential characteristics for candidates for this scholarship are:
• master degree in Electrical Engineering, Communication Engineering, Computer Science, Mathematics or equivalents;
• knowledge in pattern recognition, deep learning, image/signal processing, statistic/remote sensing, passive/active sensors.

Machine learning can impact people when it is used to automate decisions in areas such as insurance, lending, hiring, and predictive policing. In many of these scenarios, several works have shown that training data can be unfairly biased against certain populations and groups, for example those of a particular race, gender, or sexual orientation. Since training data may be biased, machine learning predictors must account for this to avoid perpetuating or creating discriminatory practices. This Phd student will work on designing and implementing innovative approaches for fair and explainable machine learning and deep learning algorithms. The selected student will have the possibility of collaborating with the activities of the Human-centric Machine Learning program of the ELLIS society (https://ellis.eu/).

End-2-End models for speech recognition have been steadily improved recently, achieving performance comparable to state-of-the-art systems. This paves the way to the adoption of such solutions also for the extraction of semantically higher information, directly from the raw speech. This allows avoiding approaches based on the combination of speech recognition followed by text processing, with consequent propagation of errors from the intermediate stages. The goal of this thesis is to develop innovative end-to-end systems, eventually based on the transformer model and sequence-to-sequence learning, to address tasks like spoken language understanding, named entity recognition, intent classification and so forth. Ideally, at the end of the doctoral thesis, the candidate would have developed a neural audio processing front end that can be applied to a variety of semantic down-stream tasks.

In the context of dialogues with chatbots it has been shown that endowing neural models with a persona profile is important to produce more coherent and meaningful conversations. Still, the representation of such personas is still very limited, usually based on simple facts. The goal of this PhD Thesis is to make a step forward, trying to grasp more profound characteristics of human personality (such as opinions, values, and beliefs) to drive language generation of conversational agents in multiple domains and languages.

Conversational agents are designed to interact with users in multiple domains on several topics using natural language. Recently end-to-end systems have started to be tested to fight fake news and hate speech in single turn settings. Still, scaling to full dialogue interactions is a challenging topic, requiring world and domain knowledge together with a deep understanding of argumentative tactics. The goal of this PhD Thesis is to overcome the shortcomings of traditional end-to-end applications in which all components are trained from the dialogs themselves, by incorporating several dialogue, argumentation and domain features.

Applications relying on the most modern sensing devices and technologies, also combining complex artificial intelligence tasks are now mainstream. The typical approach to enable intelligent applications is cloud-centric, meaning that the intelligence is hosted in the cloud infrastructure, the sensor data collected by some IoT devices. Shifting intelligence from the cloud to the edge of the network can offer different advantages such as reducing the required bandwidth and latency and also improving users’ privacy. However, reconfigure and deploy an end-to-end processing flow that involves the three aforementioned architectural layers (the cloud, the edge and embedded devices) poses major challenges: many different constraints and trade-offs must be addressed (latency, response time , bandwidth, energy consumption, computational power, computational precision, etc.) The subject of this Ph.D. is to investigate and propose novel optimization (such as e.g. pareto-based optimization) and assessment methodologies to efficiently sample such a complex design space in target application sectors such as home, industry, manufacturing, farming, etc. The reference technological environment covers (but are not limited to) embedded device software engineering (micropython, mbed OS, C languages and dialects, etc.), machine learning frameworks deployable on tiny devices (tinyML, tensorFlow lite, etc.), edge-based frameworks (eclipse Kura, edgeX Fundry, etc.) and cloud-based IoT platforms and services with AI support and components (MS Azure, AWS Greengrass, ThingsBoard, etc.).

Supervised learning is a very popular mechanism to teach machines vision-based tasks and skills. Supervision, however, is a bottleneck for building generic machines that can operate across different contexts, environments and applications, while learning and improving their understanding seamlessly. Ideally, machines should develop their own creative strategies for using the sensed data and their experience to continually learn without humans at their side.
The research activities related to this PhD position will focus on building machine vision algorithms to teach machines to seamlessly understand environments: by exploiting as little supervision as possible, by being independent of the sensor modality being used, and by updating their knowledge when ground truth information becomes incrementally available over time.

The Internet of Things (IoT), including smart objects, wearables, and wireless sensor networks, is becoming a key technology to enable applications and services in several domains. Ultra-low-power embedded devices are pervasive; novel embedded machine learning frameworks have been introduced. Thus, distributing intelligence at the edge is possible, opening exciting research scenarios spanning from novel, innovative hardware for always-on or event-based sensing up to deep learning solutions, federated learning, and continual learning fitting resource-constrained platforms.
Motivated by the challenges of these research scenarios, the research aims to (i) define novel hardware/software approaches to optimize AI at the very edge on energy-efficient embedded devices, in particular for audio processing and/or computer vision; (ii) to explore the potential of distributing and fuse the intelligence in heterogeneous nodes of an IoT (iii) to demonstrate the advantages of the investigated approaches in real-world application scenarios, such as those of smart cities.

Several important problems in modern society, such as pollution and global warming, arise from the inability to achieve cooperation between individuals over a large scale. Recent research is providing a growing evidence of the power of social influence (i.e. peer pressure), in promoting cooperative behavior. This PhD has the goal of developing computational models for modeling human behavior and social interactions and of designing data-driven strategies and incentive schemes for promoting collaboration and cooperation. These approaches will be also compared with gamification strategies in the context of real-world experiments. The work of the student will be in collaboration between two research units of Fondazione Bruno Kessler, MobS (directed by Bruno Lepri) and MoDiS (directed by Annapaola Marconi).

Machine learning and deep neural networks are extensively and successfully used to process audio on powerful computers, while several problems still need to be solved for porting the technology on low consumption devices with limited resources (both in terms of computation power and memory size).
Research is necessary to reduce the redundancy in neural models to make them portable into the internet of things framework. Along this line of research, the Ph.D. thesis will address the problem of end-to-end neural processing for audio classification, keywords spotting, and privacy-preserving audio processing on resource-constrained embedded devices, considering the trade-off between performance and energy efficiency. Advanced explorative research directions will consider how adapting continual learning techniques to low-power end-devices and if approaches such as collaborative machine-learning without centralized training data (i.e. federated learning) can help in privacy-preserving resource-constrained scenarios.

The need to translate audio input from one language into text in a target language has dramatically increased in the last few years with the growth of audiovisual content freely available on the Web. Current speech translation (ST) systems are now required to be flexible and robust enough to operate in different application scenarios and diverse working conditions. On one side, the industry calls for key features like real-time processing, domain adaptability, extended language coverage and the capability to adhere to application-specific constraints (e.g. length or lip-synch constraints in the subtitling and dubbing scenarios). On the other side, the society calls for new efforts towards inclusiveness with respect to specific categories and groups (e.g. gender-sensitivity, customization to the needs of impaired users). Both dimensions (industry and society) face the orthogonal challenges posed by the variability of audio conditions (e.g. background noise, strong speakers’ accent, overlapping speakers). The objective of this PhD is to advance the state of the art in speech translation to make ST flexible and robust to these and other factors.

We are looking for candidates willing to develop novel methodologies based on machine learning, deep learning, pattern recognition and artificial intelligence for information extraction, classification, target detection and change detection in radar and radar sounder images.
The PhD activity will be developed in the context of European Space Agency (ESA) space mission JUpiter ICy moons Explorer (JUICE) in the Jovian system. The candidate will be requested to deal with images acquired from active radar systems including Synthetic Aperture Radar (SAR) images and sub-surface radar sounding data from airborne Earth Observation missions and satellite planetary exploration missions. The activity amins in improving the understanding of subsurface structure and their impact on planetary body climate.
Besides the requirements established by the rules of the ICT school, preferential characteristics for candidates for this scholarship are:
• master degree in Electrical Engineering, Communication Engineering, Computer Science, Mathematics or equivalents;
• knowledge in pattern recognition, deep learning, image/signal processing, statistic/remote sensing/radar.

Gamification principles have proven to be very effective in motivating target users in keeping their engagement within everyday challenges, including dedication to education, use of public transportation, adoption of healthy habits, and so forth. School closures due to the COVID-19 pandemic and thus the sudden change in the management of the students' educational pathways has opened up the need for methods and digital systems able to support teachers in defining educational content and objectives for their classrooms and to keep students engaged in their training path. The goal of this PhD Thesis is to investigate approaches, techniques and tools to design and release educational digital systems for personalized and cooperative learning plans. This will be done exploiting AI techniques for adaptive gamification and will support teachers in the process of defining and monitoring dedicated learning plans for their students. At the same time, it will facilitate learning, will encourage motivation and engagement, will improve student’s participation and cooperation, and will stimulate students to expand their knowledge through dedicated learning plans and personalized feedback.

The Limadou project gathers some Italian institutions participating in the China Seismo Electromagnetic Satellite (CSES) mission. CSES consists of a constellation of satellites, designed to pursue the deepest campaign of observation of the ionosphere. One of the most important scientific goals of the mission is to look for correlations between transient phenomena in the ionosphere and seismic events. Among payloads, a set of particle detectors is devoted to the detection of charged particles trapped in the Van Allen Belts, to monitor the solar activity and to measure galactic cosmic rays of very low energy. The APP group of the Physics Department in Trento looks for candidates to a PhD programme on the analysis of the scientific data from the payloads on board the CSES-01 and those to be launched on board the satellite CSES-02 in 2022. The student will focus on time-series analyses and participate in the development of the event reconstruction software. These studies will be carried out using the most modern machine learning techniques for clustering and anomaly detection, using full information from CSES payloads. The activity will be carried out in collaboration with INFN-TIFPA, Fondazione Bruno Kessler and the Institute of the High Energy Physics of Beijing. Candidates familiar with the experimental techniques for the detection of charged particles in space are welcome, as well as basic knowledge of Machine Learning/Deep Learning is recommended.

Quantum Chromodynamics (QCD) is the theory of the strong nuclear forces. At ultrarelativistic energies the degrees of freedom are quarks and gluons and their interactions can be calculated with weak coupling methods. For sufficiently high energies, the gluon density becomes large leading to strong non-linear effects whose description is the goal of the Color Glass Condensate (CGC) effective theory. It is important to apply the latter for studying observables in the forthcoming Electron Ion Collider (EIC).

We propose to investigate the optimization of quantum simulations on trapped-ion quantum processors. The Ph.D. candidate will explore the use of quantum optimal control techniques to tailor ‘analog’ gates at the laser pulse level, as well as the optimization of ‘digital’ quantum circuits built on predetermined primitive gates. The study will identify the most effective methodology to translate near-term trapped-ion quantum computing into meaningful quantum simulations of microscopic systems.

Neuroimaging methods, like structural, functional, and diffusion MRI as well as MEG/EEG can be used to investigate the anatomical and functional connectivity of the brain. In this project, the candidate will pursue research on machine learning methods for neuroimaging data to study and characterize brain connectivity, with applications to longitudinal studies and clinical practice. The ideal candidate should have a mixed background in neuroimaging techniques and numerate disciplines, like computer science, engineering, physics, or mathematics. This project is in collaboration with the Division of Neurosurgery, S. Chiara Hospital, Trento (IT).

Deep neural networks (DNNs) have reached a prominent position amongst machine learning systems [4, 1] due to an increasing experimental evidence of their flexibility, expressivity, and effectiveness in addressing functional approximation problems. At the same time, a complete and satisfactory mathematical theory, explaining for instance how to optimally design and train a DNN on some specific task, is largely missing. It is worth noticing that typical DNNs need millions or billions of trainable parameters for achieving such outstanding performances, and must execute a proportional amount of arithmetic operations during the inference step. These characteristics make DNNs very demanding in terms of storage, memory, and energy consumption. This creates an obstacle towards their deployment on low-memory and low-power architectures such as embedded devices and microcontroller units (MCUs). For this reason, the use of quantized neural networks (QNNs), that is, of specific DNNs whose parameters take values in small, finite sets and whose activation functions have a finite range, allows to represent the operands with fewer bits with respect to standard DNNs, thus providing significant benefits with respect to digital hardware constraints [10, 5]. In this sense, QNNs have an enormous potential of applications, for instance in biomedical and environmental research, where in-situ analyses of real-time data could be effectively performed by low-powered & portable devices. At the same time, the lack of a rigorous mathematical theory in support of QNNs is even more striking, also due to the non-differentiability properties of such networks, so that a better comprehension of the mathematics involved seems crucial for their effective application. The objectives of the PhD project will be of both theoretical and applicative nature. The theoretical part of the project will consist in studying some mathematical aspects of the theory of DNNs, including: • techniques for layer-wise regularization & approximation of QNNs; • expressivity properties of suitable classes of QNNs;

The Ph.D. student will evaluate control measures against different infectious diseases, both retrospectively and prospectively, by developing mathematical models calibrated against observed epidemiological data and informed by other data relevant to the infection under study. The modelling approach will be tailored to the addressed problems and may include compartmental models, generative models, individual-based simulations as well as Bayesian approaches and will include scenario analysis to compare alternative intervention strategies.

Research activity conducted during the Ph.D. will focus on the development of epidemiological models informed by real-world data aimed at investigating the main determinants of the disease spread in humans. Envisioned approaches range from the study of mechanistic models mimicking the spatio-temporal transmission dynamics of infectious diseases to the use of bayesian approaches applied to detailed epidemiological records.

Interpretability has become a crucial requirement to support translation of reliable AI models into clinical practice. Throughout the graduate course, the candidate will explore how to make a DL model interpretable and reproducible, and she/he will test such methodologies comparing predictive models trained on integrated imaging (CT/PET/MRI or Digital Pathology), omics and clinic data, both publicly available and original, with the final goal of defining a framework or a pipeline leading from the available data to the explainable and repeatable models.

Redox flow batteries (RFBs) are a promising technology for large scale energy storage. In RFBs power and energy are decoupled: the former depends mainly on the size of the stack while the latter on the size tanks containing the redox active species. This feature make RFBs ideal for economical, large- scale energy storage. However, cost reduction are in order for allowing a widespread diffusion of this technology. The required cost reductions involve two main components of the system: the electrolytes and the stack. Both need to be optimized for enabling a large scale diffusion of RFBs.
Flow batteries are a complex system their design and optimization usually leads to a trade-off between cost and performances (energy and power density, efficiency, cycling life). Cell and stack design is a core task required for the development and upscaling of flow battery systems but redox flow cells models can be very complex due to the multitude of physical phenomena that need to be considered: electric fields, fluid flow, mass and specie transport in different components, electrochemical reactions, heat transfer. All these phenomena need to be considered for building a digital twin of cell and stack and enable to identifying cell-limiting mechanisms, forecasting cell performance and optimizing the design. Despite some commercial software (e.g., COMSOL) can support this activity these tools present a lack of flexibility and serious constraints concerning their use on HPC platform as well as their performances (this also limit their use to small scale cells). Furthermore multiscale models that couple detailed cell models and system level models are not currently available.
For the above mentioned reasons, in this PhD topic we propose to develop a multiphysics-multiscale tool aimed at supporting redox flow cell and stack design and upscaling. This tool will also enable design optimization supported by different algorithms. The selected candidate will be in charge of developing the models extending opensource modelling platforms such as OpenFOAM and integrating optimization tools such as Dakota. The platform will be composed of three different main components tightly connected with each other: 1) Multiphysics cell and stack model (using for example OpenFOAM), 2) System level model based on transient 1D-0D models (using OpenFOAM, OpenModelica or python) 3) optimization tool. The outcome from the multiphysics cell model will be used either as input or for computing the parameters required by the system level model. Different type of optimization will be developed based on the final objective.
The simulation models will be validated with experimental data from known chemistries and representative prototypes, and show how new chemistries can be explored. The candidate will be in charge of developing and implementing the physical models, validating the models based on experimental data, integrating different model for building a multiscale tool and integrating the optimization algorithms in the work flow to enable design optimization. In order to enable a strong cross-contamination of ideas and experience, we propose that the candidate will also support the experimental activities related to the validation of redox flow cells with known and new chemistries.

In the last years, mobile mapping systems, such as hand-held devices, drones and ground vehicles equipped with active or passive sensors, have been widely used for precise 3D data generation based on advanced geo-referencing solutions.

According to specific scenario and application requirements and constraints, different techniques can be adopted to tackle positioning and navigation tasks, e.g. solutions based on global navigation satellite system (GNSS), ultra-wide band (UWB) transceivers and 5G mobile network technologies.

The research should investigate different geo-referencing solutions when using mapping platforms, trying to understand potentials and limitations, feasibility and needs in typical geomatics scenarios, such as mapping of urban and forestry areas, precision farming, hazard monitoring, indoor mapping, heritage documentation.

Some topics that should be considered in the research are:

- the forthcoming Galileo high accuracy service, which should be investigated and exploited at the best of it state of implementation;

- the fusion of data from multi-frequency and multi-constellation GNSS receivers and AI-enabled stereo cameras or stereo optical and depth sensors embedded in smart phones;

- alternative positioning solutions such as those based on UWB and 5G technologies.

The mentioned topics should be investigated in particular for the evaluation of GNSS- based positioning performances and impact on the 3D data in mobile mapping applications, in particular in the photogrammetric field. Positioning technologies such as RTK, PPP and PPP-RTK should be considered along with suitable strategies for the mitigation of GNSS signal degradation and complete temporary loss.

The research is expected to advance the knowledge on the impact of most recent geo-referencing solutions in geomatic mapping based on the photogrammetric principles. High-accuracy positioning is expected to improve the quality of 3D geometric and spatial data and modelling to be used in professional applications. The output of the research will be made accessible through scientific papers.

Computational thinking together with digital competences and AI education are emerging as important topics in the field of education and technology enhanced learning. Although several courses and educational material are being developed, there is still a lack of technology-based personalized and inclusive approaches that might be used to improve teaching and learning practices for such concepts.The aim of this PhD project is to investigate new technology-based approaches to develop computational thinking skills, to improve digital competences and to make AI education accessible. These may include (but not limited to) end-user programming for teachers and students for personalizing and co-creating learning activities. The ideal candidate has a background in Computer Science, Psychology or Cognitive Science. Experience with design of interactive digital technologies, conduction of experimental and in-the-wild studies as well as competences with educational theories are a plus for the application and should be acquired during the Phd training. The PhD position is offered in co-tutoring between the i3 research unit of the Digital Society center of FBK and the Department of Psychology and Cognitive Science.

In recent years there has been a growing interest for psychoeducational interventions delivered by means of mobile applications and personal assistants to support self-care of patients, including those coping with chronic conditions. Although the validity of psychoeducation has been proved repeatedly by previous research, the design of effective behavioral intervention technologies for virtual coaching in the area of healthy coping remains a challenge. The aim of the PhD project is to investigate key features of smart coaching solutions for healthy coping interventions that are engaging to use by patients and produce effective outcomes from a clinical perspective. The ideal candidate will be strongly motivated in developing design skills in the field of behavioral intervention technologies and conversational agents for applications in healthcare. The PhD position is offered in co-tutoring between the Digital Health Lab of FBK and DIPSCO.

Game-based motivational systems are emerging as an effective tool to engage users and induce a positive change in human behavior. Games introduce goals, interaction, feedback, problem solving, competition, narrative, and fun learning environments, motivational affordances that can increase end-user engagement and motivation. Gamification has gained significant attention especially in educational contexts, where supporting and retaining students motivation is a constant challenge. Although research has demonstrated that collaborative learning benefits a variety of learning outcomes, while also supporting people’s social, emotional, and psychological well-being, most studies on gamification in the educational context focus on individual work and competitive learning activities, exploiting affordances that highlight each student achievements and progression. The goal of this PhD thesis is to investigate the potential of combining motivational gamification mechanics and social and interactive elements of collaborative learning, analyzing the impact in terms of
students’ achievements, engagement, participation and cooperation. The ideal candidate has a background in Computer Science or Cognitive Science. Game design, educational and cognitive psychology, motivation theories, knowledge on designing and conducting experimental studies, experience with quantitative and qualitative data analysis techniques are a plus for the application and should be acquired during the Phd training.

The Phd candidate will deal with the analysis of olfactory information using digital methods. In particular, s/he will analyse how odors are described in multiple languages, and how smell-related terminology has evolved over time. S/he will also contribute to the development of a system to extract olfactory information from texts, and to the linguistic analysis and evaluation of the system output.

In recent years there has been a growing interest in conversational AI, and a number of conversational systems are now operative in various sectors, including call centres. This situation has made available a huge amount of user-machine interactions, which have a high potential to be used to improve the system performance. As an example, the capacity to detect the emotional content of the conversation would allow the system to respond in a more appropriate way to the user requests. This PhD grant addresses some of the scientific challenges which are behind the interpretation of very large-scale conversational data, including managing noisy data, topic clustering, semi-supervised intent classification, and emotion detection. The resultant of these techniques will be applied to develop empathic chatbots able to model their answers, in real time, on the basis of the human detected emotions expressed during the conversations.

Required/Preferred Candidate Skills and Competencies: Required: Master’s Degree in Science, Computing & Technology, Statistics, Engineering or Mathematics Preferred: Documented experiences in the use Machine Learning techniques applied to real data.

The intellectual property of the research results that will derive from the activities carried out by the doctoral student is owned by the Company.

Some recent works have shown that the combination of high-resolution satellite imagery and machine learning/deep learning techniques have proven useful for a range of socio-economic and sustainability-related tasks, from poverty prediction to population mapping, from forest and water quality monitoring to the mapping of informal economic activities and settlements, etc. This PhD project aims at designing novel deep learning approaches and novel ways for combining satellite imagery collected at different temporal and spatial resolutions, combining different types of data (for example, optical + radar), and/or combining satellite imagery with other relevant data such as information captured by mobile phones. Moreover, special focus will be dedicated to address applications characterized by the limited amount of reference training information (e.g., property valuation, spatial wealth distribution, exposure to respiratory diseases, etc.). The ideal candidate is strongly motivated to develop machine learning and remote sensing skills focusing on deep learning, satellite imagery and multi-modal approaches, as well as interested in applications to socio-economic and sustainability challenges. The project will be supervised by Bruno Lepri (FBK), Emanuele Strano and Mattia Marconcini (MindEarth).

Required/Preferred Candidate Skills and Competencies: Background in computer vision with knowledge on object recognition, image segmentation, deep learning techniques. Interest for applications to visual scene understanding (in particular, urban environments).

This PhD project is focused on using deep learning approaches, computer vision and photogrammetry for understanding visual scenes, in particular related to urban environments and people’s behaviours. The project aims at designing novel deep neural network architectures able to exploit multiple sources of data efficiently and to detect people's behaviours, objects, vehicles, in crowded environments such as streets, squares, malls, etc. For example, the project might involve street-view imagery in conjunction with satellite imagery, 3D data, etc. to predict urban outcomes and people’s behaviour. In addition, the candidate will work on generative models (e.g. Generative Adversarial Networks) to augment training data on urban and behavioral patterns (e.g. people movements), which have to be realistic and capture the high diversity of urban forms and lifestyles observed across the globe. Special emphasis will be given to exploit synthetically generated data within GAN architectures. The ideal candidate is strongly motivated to develop machine learning skills focusing on deep learning computer vision and multi-modal approaches. The project will be supervised by Bruno Lepri (FBK), Nicu Sebe (DISI), and Emanuele Strano (MindEarth).

Required/Preferred Candidate Skills and Competencies: Background in computer vision with knowledge on object recognition, image segmentation, deep learning techniques. Interest for applications to visual scene understanding (in particular, urban environments).

Hydrogen is the most promising among the potential green gases, an essential energy carrier to enable a deep decarbonization, for the sectors difficult to abate, such as heavy industry and heavy mobility. Hydrogen indeed must be extracted by water through electrolysis or other materials of biological origin and wastes. One of the most urgent needs to solve is the scaling up of the sector involving several solutions in the way hydrogen is stored, moved, transported in between production and utilization. Beyond compression, one promising direction for some sectors is that of energy carriers, such as liquid hydrogen and liquid organic hydrogen carriers. Among these, ammonia. For its specific characteristics, ammonia could be an ideal carrier in terms of physical and chemical properties, energy density, enabling an efficient logistic and an ideal use in the hydrogen chain. One of the gaps is its synthesis and its direct utilization. This is potentially feasible through innovative technologies, such as reversible Solid Oxide Cells. The PhD will focus on these two dimensions to enable a safe and efficient generation of ammonia in (co)-electrolysis processes through Solid-State Ammonia Synthesis (SSAS) and its utilization in Direct Ammonia Fuel Cells The PhD will focus on both modelling and engineering as well as on experimental and validation activities for cell and short stack based Solid State technology. The activities will include: • Preliminary study on enabling key technologies for the ammonia synthesis; • Engineering study for direct ammonia synthesis, to design and develop both the single components level and the overall integrated system in terms of sizing, Balance of Plant, integration layout and controls; • Demonstration on lab scale of ammonia synthesis through SSAS process and its utilization in Direct Ammonia Fuel Cells based on Solid oxide and/or Proton Conductive Ceramic technologies.

Required/Preferred Candidate Skills and Competencies:
- Competences in energy engineering;
- Knowledge of the Hydrogen chain;
- Knowledge on conversion processes using both Electrolyser and fuel cells;
- Lab training in use of hydrogen related compounds, including hydrogen carriers.

Among the low temperature electrolysis processes, two main approaches are extensively documented: alkaline water electrolysis (AWE) and proton exchange membrane electrolysis cell (PEMWE). AWE is a well-established and durable technology yet with many shortcomings being a large footprint, difficulties in handling the liquid alkaline electrolyte, and insufficient response time. Anion exchange membrane water electrolysis (AEMWE) can potentially combine the beneficial features of the PEMWE and AWE technologies, low cost, raw materials that do not raise concerns in terms of supply bottlenecks, electrodes that do not include platinum group metals (PGM), stainless steel porous transport layers (PTL) and bipolar plates (BPP), a compact design, the adoption of feeds based on noncorrosive liquids (low concentration alkali or pure water), and differential pressure operation. However, as of today AEMWE is limited by AEMs exhibiting an insufficient ionic conductivity as well as a poor chemical and thermal stability. The thesis will focus onto the development and testing of innovative cell layout and materials and in parallel on the improvement/optimization of existing AEM WEL cells. A second aspect of the work is the development of a novel concept stack AEMWE, based different geometry of flow field and electrolyte distribution to extend the dynamic range of operation as well as the reduction of gas cross over to achieve an high purity hydrogen output. The key concept is to reduce the voltage and increase the current density. The approach of the work will articulate around the following aspects: design of high surface area cell, developing suitable support material for this approach which needs to be active towards water dissociation, selection of the more performant components and finally validation of the AEM-WEL stack layut best candidate.

Required/Preferred Candidate Skills and Competencies: consistent, diligent, independent, innovative, good command of English. Knowledge of Italian is strongly preferred.

AEM (Anionic Exchange Membrane) technology represents one of the most advanced and promising technology for the low temperature electrolysis process for the production of green hydrogen. The most important advantage is given by the low cost, if compared with other similar technologies, like the PEM based electrolysis. The cost reduction is given mainly by the adoption of non-precious-metal compounds in the catalytic layers. Currently the AEM technology is not fully mature for large commercial applications, due to the gap of the AEM durability and efficiency, compared to the other, mature technologies. The main target of this research activity will be focused on the development of advanced deposition technologies to improve the AEM performances in terms of efficiency and durability. More in particular, the research activity will be focused on electrospinning coatings of catalytic nanofibers and on spray coatings technologies. Dry synthesis technologies based on the use of plasma and chemical-plasma hybrid processes will be applied too, as well as sophisticated surface analysis techniques. Physical vapor deposition, surface plasma functionalization and atomic layer deposition techniques will be employed for the membrane surface and bulk properties modification following as much as possible on-pot and scalable methods. X-ray photoelectron and Auger electron spectroscopies will be used to investigate the chemical properties of the membrane’s surfaces related to the searched functionalities, along with the physical properties and mechanical stability of the membranes which will be studied by means of electrical, mechanical and stress measurements. AEM will be characterized about V-I behaviour and impedance measurements to investigate limiting transfer mechanism and main barriers. The final target and expected outcome will also include a preliminary analysis of the scale up of the coating technology to a mass production level, for commercial electrolyzer applications.

Required/Preferred Candidate Skills and Competencies:
- Catalytic materials;
- Advanced nanomaterials for energy applications;
- Deposition techniques for nanostructured layers and thin films;
- Electrolysis technologies.

A clinical case is a statement of a clinical practice, presenting the reason for a clinical visit, the description of physical exams, and the assessment of the patient’s situation. Clinical cases (e.g. discharge summaries, clinical cases published in journals, and clinical cases from medical training resources) provide a very valuable source of information for data-driven technologies aiming at predicting clinical outcomes and patient behaviors. This three-year PhD offers a unique context of a collaboration between FBK, specifically the NLP group, and a Swiss multinational healthcare company, worldwide leader in biomedical research. The research will focus on a number of application oriented tasks, including automatic recognition of clinical entities (e.g. pathologies, symptoms, procedures, and body parts, according to standard clinical taxonomies such as ICD-9, ICD-10 and SNOMED-CT); detection of temporal information (i.e. events, time expressions and temporal relations, according to the THYME TimeML standard), and factuality information (e.g. event factuality values, assessment of the effect of negation, uncertainty and hedge expressions). Italian will be the major language of clinical cases, although technologies will be experimented on other languages. The goal of the PhD is both to advance the state of the art for clinical case analysis for the Italian language, and to deliver prototype applications, which can be further made operative in real settings (e.g. hospitals). The candidate will have the unique opportunity to explore different domains (Natural Language Processing, Machine Learning, Health & Well-Being) being directly coached by very experienced teammates. The involved PhD will work in an international environment, collaborating with a healthcare company, with worldwide presence. The candidate will work both at FBK (Trento) and at the abovementioned company’s premises (both in Italy and abroad).

Required/Preferred Candidate Skills and Competencies: the candidate should possess basic knowledge on Natural Language Processing and Machine Learning techniques (particularly deep learning architectures). Experience on biomedical data will be a plus. Basic programming skills (e.g. Python) would complete the profile. Proficiency in English is required, basic knowledge of Italian preferable.

Dialogue agents usually work in limited domains with clear and well defined policies, with little adaptation capabilities
to the contextual and social situations. In this scenario it has been shown that endowing neural models with a consistent user/machine
profile is important to produce more coherent and natural conversations. Still, the representation of such profiles is very limited, usually based on simple facts. The goal of this PhD Thesis is investigating end-to-end approaches and generation strategies such as guided decoding, in combination with large pre-trained language models and external knowledge, in order to improve the naturalness of
dialogues. During the PhD we will always be mindful of the societal impact of the technologies we develop.

In recent years there has been a growing interest for psychoeducational interventions delivered by means of mobile applications and personal assistants to support self-care of patients, including those coping with chronic conditions. Although the validity of psychoeducation has been proved repeatedly by previous research, the design of effective behavioral intervention technologies for virtual coaching in the area of healthy coping remains a challenge. The aim of the PhD project is to investigate key features of smart coaching solutions for healthy coping interventions that are engaging to use by patients and produce effective outcomes from a clinical perspective. The ideal candidate will be strongly motivated in developing design skills in the field of behavioral intervention technologies and conversational agents for applications in healthcare. The PhD position is offered in co-tutoring between the Digital Health Lab of FBK and UNIPD.

With this phd, we want to investigate how recommendation systems can take advantage of integrating background knowledge expressed in some symbolic form as e.g. logical formulas with standard numeric optimization method, based on machine learning approaches.
In particular we want to investigate how the state-of-the art methods for recommended systems based on embeddings can be complemented with logical reasoning on formulas that expresses structural properties and constraints about properties and relations between users and items.

Machine learning and deep neural networks are extensively and successfully used to process audio on powerful computers, while several problems still need to be solved for porting the technology on low consumption devices with limited resources (both in terms of computation power and memory size).
Research is necessary to reduce the redundancy in neural models to make them portable into the internet of things framework. Along this line of research, the Ph.D. thesis will address the problem of end-to-end neural processing for audio classification, keywords spotting, and privacy-preserving audio processing on resource-constrained embedded devices, considering the trade-off between performance and energy efficiency. Advanced explorative research directions will consider how adapting continual learning techniques to low-power end-devices and if approaches such as collaborative machine-learning without centralized training data (i.e. federated learning) can help in privacy-preserving resource-constrained scenarios.

The Internet of Things (IoT), including smart objects, wearables, and wireless sensor networks, is becoming a key technology to enable applications and services in several domains. Ultra-low-power embedded devices are pervasive; novel embedded machine learning frameworks have been introduced. Thus, distributing intelligence at the edge is possible, opening exciting research scenarios spanning from novel, innovative hardware for always-on or event-based sensing up to deep learning solutions, federated learning, and continual learning fitting resource-constrained platforms.
Motivated by the challenges of these research scenarios, the research aims to (i) define novel hardware/software approaches to optimize AI at the very edge on energy-efficient embedded devices, in particular for audio processing and/or computer vision; (ii) to explore the potential of distributing and fuse the intelligence in heterogeneous nodes of an IoT (iii) to demonstrate the advantages of the investigated approaches in real-world application scenarios, such as those of smart cities.

As demonstrated by recent events, telecommunications networks’ importance for economic activity and simple human communication cannot be understated. However, while networks held up remarkably well despite a near-doubling of their expected load, their resiliency is not infinite, and their role as information carriers makes them useful targets for malicious actors, from crooks to state-sponsored agents. The advent of programmable ASICs offers a unique opportunity for researchers to observe and customize network device behaviors at a level of detail and time resolution unthinkable with traditional approaches. These, in turn, enable the development and application of new or previously unsuitable strategies for on-the-fly fault detection, isolation and recovery policies, including aspects such as detailed timing and direction of error propagation and intrusion detection and isolation.
This thesis combines the topics of programmable networks and advanced fault and intrusion detections and recovery/isolation, with the aim of improving the resiliency of large-scale telecommunications networks against both failures and targeted attacks.

Machine Learning (ML) is nowadays a consolidated technology embedded in various domains of computer science and information technology. In the recent past years, ML has revolutionised cybersecurity applications, with excellent results in various application areas such as: encrypted traffic classification, intrusion detection and prevention, anomaly detection in industrial control systems, identification of malicious software (or malware), among others.
One important research subfield of ML is called Explainable Machine Learning, which relates to understanding the ML model behaviour by means of various techniques such as feature importance scores, influential training data, etc,. Given the complexity of some black-box ML models, it is inherently difficult to understand why they behave the way they do.
Understanding how a ML model works and how it takes its decisions is paramount in network security. Indeed, the ability to understand why an event is classified as benign or malicious by an ML-based intrusion detection system allows the ML practitioner to take the necessary counteractions to reduce false positive and false negatives rates, and to make the system more robust to Adversarial Machine Learning attacks.

The objective of this thesis is to perform fundamental research in the field of ML explainability (understanding how ML algorithms reason their outputs) and to propose novel tools and methodologies for ensuring good performance of ML-based security systems under various working conditions.

Space and avionics systems are reaching an unprecedented degree of complexity. The process of safety analysis attempts to characterize the likelihood of faults and failures, and to assess the effectiveness of the adopted mitigation measures. Unfortunately, traditional techniques are becoming ineffective, unable to deal with large-scale systems. This thesis will investigate novel methods for safety analysis, based on the adoption of formal models of system and software (nominal and faulty) behaviors. Particularly interesting are the analysis of timing aspects in the propagation of multiple faults to failures and errors, the ability to explain the causality of propagation, and the definition of techniques for on-the-fly fault detection, isolation and recovery policies.

The PhD project aims at the investigation of innovative functional nanomaterials towards the selective detection of gaseous compounds. The combined use of microfabrication process, silicon functionalization and solid-state gas sensor technology enables a new breakthrough approach in the development of high-performing gas sensing devices, useful for different applications such as indoor and outdoor air quality monitoring, precision farming and medical screening.

Advanced nanostructured materials can be exploited both to improve the gas sensor performance, and to develop innovative gas monitoring tool, such as monolithic micro-gas chromatographs. These innovative gas monitoring systems requires an interdisciplinary investigation into the operating principles of the gas chromatographic technique. The synergistic effect of the system fluid dynamics, combined with the chemistry of surface interactions, represents the basis of this technology. In particular, the chemistry of heterogeneous solid-gas and/or liquid-gas interactions play a key role in the separation and detection of the analyzed molecules. The adequate functionalization with polar or non-polar stationary phases of the pre-concentrator and of the chromatographic microcolumn, developed by means of silicon microfabrication techniques, is crucial to obtain an adequate separation of the analytes. Likewise, the development and deposition of specific nanostructured materials, which act as an active sensing layer in the sensing platform, enables the detection of previously separated compounds

This PhD grant will exploit cross-disciplinary competences in three areas, i.e., deep learning, argumentation mining and conversational AI, to support a broader and innovative view of explainable AI. The goal is to advance the state of the art in explanatory dialogues through the capacity to automatically detect the quality of an argument. The grant is related to the ANTIDOTE project (Argumentation-Driven explainable artificial intelligence for digital medicine), which fosters an integrated vision of explainable AI, where low level characteristics of the deep learning process are combined with higher level schemas proper of the human argumentation capacity. The research will focus on a number of deep learning tasks in the medical domain, where the need for high quality explanations for clinical cases deliberation is more critical than in other domains.

The aim of this thesis is investigating how to develop, exploit, and combine techniques and approaches borrowed from different research fields, ranging from logic to artificial intelligence, from model checking to statistics, to advance the existing services for process mining.
To this purpose, several are the challenges to be faced in the work as, for example, (i) the capability to represent and exploit secondary aspects for business processes such as data, time, resources; (ii) the capability to align execution information with models; (iii) the capability to manage and reason on extremely large quantity of data (big data);

In knowledge-intensive Digital Society applications it is essential to be able to integrate heterogeneous, large and uncertain knowledge sources and infer new knowledge emerging from such an integration process. We are interested in developing new methods for tackling these problems, both from the foundational and the practical point of view. We identify two possible directions: 1) Ontology-mediated transformation of knowledge structures for efficient migration
We aim at developing solutions for migration of data between heterogeneous representation schemas: an ontology is used as an abstract model of the domain of data; ontology aligned schemas will provide data access as a Virtual Knowledge Graph.
We are interested in automatizing ontology mappings from/to resources and logic-based generation migration rules across resources. 2) Reasoning with weighted information to handle context and exceptions. We recently considered a solution (justifiable exceptions) for reasoning over ontological knowledge that allows for exceptional instances. However, in real-world data it is often useful to quantify the degree of exceptionality or probability of an axiom. We are thus interested in studying defeasible reasoning in (contextualized) description logics by integrating numerical reasoning as well as novel methods of explainability in AI.

Integrating formal verification and testing for parametric software product lines

The increasing complexity and configurability of systems requires the development of new methods and tools to design and test parametric software systems and product lines characterized by variability from the point of view of the space of the possible functional configurations, the space of the deployment architectures, and of the aspects related to their dynamic reconfiguration at run-time. This Ph.D. thesis aims at defining novel approaches to the testing, verification and validation of this class of systems exploiting the integration of formal verification techniques and software testing approaches. On the one hand, formal techniques have the potential to analyze and cover all the behaviors of a software system; on the other hand, testing techniques are able to ensure adequate levels of coverage also in case of large and complex systems. The challenge here is the search for new directions and methods to interleave the two approaches in the specific case of large parametric systems, in order to optimize the balance between the adequacy of the coverage of the system's behaviors and the effort of the overall verification activity.

Reconfigurable and trustworthy pandemic simulation

Simulation tools are fundamental to predict the evolution of pandemic, and to assess the quality of counter-measures, e.g. the effect of travel restrictions on the spread of the coronavirus. However, they come with two fundamental requirements. The first is the need for a fast reconfiguration of the simulation, in order to be able to describe the mutating scenarios of the pandemics. The second is the ability to produce correct and explainable results, so that they can be trusted and independently validated. The topic of this research is to devise a model-based approach that is able to represent at a high-level the features of a generic pandemic, from which an efficient simulator can be produced. By means of formal methods, the results of the simulation are guaranteed to be correct by construction, with proofs that can be properly visualized and independently checked. The activity will be carried out as a collaboration of the Center for Health Emergencies (https://www.fbk.eu/it/health-emergencies/), that played a major role during the ongoing pandemics, and the Center for Digital Industry (https://dicenter.fbk.eu/), a leading centers in model-based design.

Condition monitoring and predictive maintenance of complex industrial systems: Model-based reasoning meets Data Science

The advent of Industry 4.0 has made it possible to collect huge quantities of data on the operation of complex systems and components, such as production plants, power stations, engines and bearings. Based on such information, deep learning techniques can be applied to assess the state of the equipment under observation, to detect if anomalous conditions have arised, and to predict the remaining useful lifetime, so that suitable maintenance actions can be planned. Unfortunately, data driven approaches often require very expensive training sessions, and may have problems in learning very rare conditions such as faults. Interestingly, the systems under inspection often come with substantial background knowledge on the structure of the design, the operation conditions, and the typical malfunctions. The goal of this PhD thesis is to empower machine learning algorithms to exploit such background knowledge, thus achieving higher levels of accuracy with less training data.

Planning and scheduling with time and resource constraints for flexible manufacturing

Many application domains require the ability to automatically generate a suitable course of actions that will achieve the desired objectives. Notable examples include the control of truck fleets for logistic problems, the organization of activities of automated production sites, or the synthesis of the missions carried out by unmanned, autonomous robots. Planning and scheduling (P&S) are fundamental research topics in Artificial Intelligence, and increasing interest is being devoted to the problem of dealing with timing and resources. In fact, plans and schedules need to satisfy complex constraints in terms of timing and resource consumption, and must be optimal or quasi-optimal with respect to given cost functions. The Ph.D. activity will concentrate on the definition of an expressive, formal framework for planning with durative actions and continuous resource consumption, and on devising efficient algorithms for resource-optimal planning. The activity will explore the application of formal methods such as model checking for infinite-state transition systems, and Satisfiability and Optimization Modulo Theories, and will focus on practical problems emerging from the flexible manufacturing domain.

Meta-learning for efficient 3D representations

Learning-based algorithms for 3D object description, recognition and retrieval suffer from lack of annotated data, computationally inefficient processing pipelines and poor generalisation ability across different application domains, such as robotic manipulation, automotive, and augmented and virtual reality. All these factors together often hinder the employment of 3D processing pipelines in real-world applications. The goal of this Ph.D. position is to conduct research on novel efficient algorithms for 3D feature representations using deep learning that can effectively replace traditional hand-crafted modules to ultimately improve performance, ease deployment and foster scalability.

Deep neural networks are being extensively and successfully used in a variety of audio processing tasks. However, neural solutions typically are computationally eager, in particular during training, and are not suitable for low consumption devices with limited resources (both in terms of computation power and memory footprint). TinyML (or tinyAI) is a branch of neural machine learning that investigates strategies to make AI portable on constrained platforms as those typically available in internet of things frameworks. Along this line of research, the Ph.D. thesis will address the problem of end-to-end neural processing for audio classification, keywords spotting, and privacy-preserving audio processing on resource-constrained embedded devices, considering the trade-off between performance and energy efficiency. Advanced explorative research directions will consider how adapting continual learning techniques to low-power end-devices and if approaches such as collaborative machine-learning without centralized training data (i.e. federated learning) can help in privacy-preserving resource-constrained scenarios.

Cloud adoption is on the rise and promises to offer many advantages by leveraging economies of scale; at the same time, new security and privacy challenges arise. As an example, consider the protection of data; while in transit and at rest, cryptographic techniques to guarantee confidentiality and integrity are well-understood and readily available for several different use case scenarios in the cloud. The situation is much less clear for data in use, i.e. during computation, although it is fundamental for trusting cloud service providers without taking for granted their unsupported claims about security assurances especially when sensitive (e.g., healthcare or financial) information is being processed. To achieve this advanced level of data protection, it is fundamental to design and prove the security of technical enforcement mechanisms of confidentiality and integrity policies in Trusted Execution Environments. For the usability of these mechanisms, fundamental security services (including key management and attestation) must be developed, their security and risk level formally assessed, and their deployment automated. The research work to be conducted during the thesis aims to make significant contributions to developing methodologies, automated techniques and tools to assist the development of fundamental services for confidential computing solutions in the cloud with a focus on key management, attestation, and their integration with identity management solutions for both users and machines to establish a root of trust with high assurance. Applications of interest for the integration of foundational services range from confidential AI, databases, and analytics to confidential ledgers and multiparty collaboration of dataset owners.

Autonomy requires the ability to devise a suitable course of actions, execute the resulting plan, monitor its execution, detect the presence of faults that may have occurred, identify their nature, recover from them, and replan. In addition to achieving the desired goals, planning is required to optimal with respect to multiple criteria, including resource consumption, timings, and tolerance to faults. The problem addressed in this thesis is to define a trustworthy architecture for autonomy, whose correct operation can be formally proved. This shall be carried out by representing planning, runtime monitoring, fault detection, identification and recovery (FDIR), and robustness evaluation in a unique logic-based framework. Satisfiability Modulo Theory will be used to represent with real-time constraints, nondeterminism, resource consumption and temporally extended goals. Within this framework, the algorithms for planning, monitoring and FDIR will be inspired by modern model checking procedures based on incremental abstraction-refinement to deal with infinite state spaces.

Chemoresistive gas sensors are high-performance sensing devices with very attractive properties, such as high sensitivity, small size and easy integration into portable eletronics and IoT platform. However, their low selectivity is limiting their full potential in many applications. In order to overcome this shortcoming, the use of artificial intellingence approaches for data data analysis is the most promising. Therefore, the adoption of specific sensor arrays equipped with dedicated algorythm turned out to be the successful strategy for the development of optimized sensing platform for medical applications and the indoor air quality monitoring. The PhD project here proposed aims at the investigation of innovative solutions for data analysis of chemoresistive gas sensors, especially for outdoor air quality monitoring and precision agriculture.

Fog computing extends cloud computing technology to the edge of the infrastructure to support dynamic computation for IoT applications. Reduced latency and location awareness in objects’ data access is attained by displacing workloads from the central cloud to edge devices, within the so-called cloud-to-edge continuum. By doing so, fog computing enables low-latency and privacy-preserving processing of data close to where it is actually produced and consumed. In addition to that, it overcomes communication bottlenecks and reduces costs, representing a key step towards the pervasive uptake of IoT/edge-based services.
This PhD project has the ambition of investigating dynamic resource orchestration methods and algorithms to smartly deploy applications in fog computing environments, with the aim of catering to the needs of applications while optimising the utilisation of the infrastructure. The work will cover both theoretical and practical aspects.

Fog Computing recently emerged as the specialisation of cloud computing to store, manage, and process information close to the edge, where data is actually produced and consumed, to support different IoT applications and use cases that are sensitive to e.g. latency, privacy, bandwidth, etc. Given the amount and heterogeneity of devices, data and applications involved, resilience to anomalies, misconfigurations and security threats is now becoming a major concern for all the involved stakeholders. In this context, Artificial Intelligence (AI) is offering powerful instruments (e.g. Deep Learning) to effectively detect attack patterns and discover threats. The goal of the proposed PhD position is to study and develop novel methods and algorithms for anomaly detection and mitigation in fog-enabled computing environments, with specific attention to balancing high detection accuracy and minimal overhead. AI and other techniques will support the dynamic detection of anomalies, the runtime execution of security functions, the rerouting of traffic and the reconfiguration of existing policies.

Machine learning and deep neural networks are extensively and successfully used to process audio on powerful computers,  while several problems still need to be solved for porting the technology on low consumption devices with limited resources  (both in terms of computation power and memory size).

Research is necessary to reduce the redundancy in neural models to make them portable into the internet of things framework. Along this line of research, the PhD thesis will address the problem of end-to-end neural processing for audio classification, keywords spotting and small vocabulary speech recognition on resource-constrained embedded devices, considering the trade-off between performance and energy efficiency.

The combination of Cloud and Edge Computing is a key-enabler for the deployment of innovative applications and services in scenarios such as cooperative, connected and automated mobility, industry 4.0, home automation and smart cities. On the positive side, the new paradigm promises to enable the exploitation of the potentially unlimited computational capabilities of the Cloud and the low-latency processing of the Edge. On the negative side, the distributed nature of Edge computing entails new security and privacy challenges. Indeed, data generated by these applications may be processed by many devices before eventually being stored in the Cloud. With the new General Data Protection Regulation (GDPR [1]), preserving data privacy is no more an option; it has become a pressing necessity. One of the consequences of this situation is the need for enforcing security (e.g., access control) and privacy (e.g., data protection) policies to fulfil increasingly complex security and privacy requirements. To make policy enforcement even more difficult, providers offering Cloud and Edge services are known to pose a threat to data privacy by actively collecting and analysing such data [2] (this is known as "Honest but Curious" providers). As a result, data must be encrypted throughout their lifecycle so that only authorised devices can actually decrypt the data. Unfortunately, cryptography is still a challenging task given the low resources deployed at the Edge (e.g., IoT devices with limited memory and computational power), and this makes the enforcement of expressive policies a difficult task.

Therefore, there is an urgent need to design innovative solutions to enforce rich security and privacy policies in resource-constrained distributed systems involving "Honest but Curious" Cloud and Edge providers. Alternative enforcement mechanisms should be deployed by considering the characteristics of the various components integrated into a Cloud-Edge system (e.g., centralized policies for the Cloud or lightweight cryptography for IoT devices). The enforcement of policies should be distributed to minimize the effort based on the available resources while still guaranteeing the completeness and efficacy of the policy at the logical level. The design of these solutions must consider usability and efficiency by adopting current best practices and using tools that assist the correct implementation.

The research work to be conducted in the thesis aims to make significant contributions to developing methodologies, techniques, and tools for the security and privacy of data in Edge-enabled applications. The activities will include:

* Analysis of the state-of-the-art solutions in securing and protecting data of Edge-enabled applications;
* Design of an innovative solution allowing to efficiently distribute the enforcement of policies in Edge-enabled applications;
* Implementation of such a solution adopting the current best practices to guarantee efficiency and frictionless user experience.

References:

[1] https://gdpr.eu/

[2] E. Ramirez, J. Brill, M.K. Ohlhausen, J.D. Wright, and T. McSweeny, "Data brokers: A call for transparency and accountability". In Data brokers: A call for transparency and accountability, pages 1–101. CreateSpace Independent Publishing Platform, January 2014.

A large number of data breaches are caused by passwords, that represent the weakest link in the authentication chain since they are usually pretty simple to guess and reused across many different online accounts. Therefore, authentication strategies are moving towards more secure solutions as Multi-Factor Authentication (MFA) [1] and Passwordless Authentication [2] (e.g., using push notification or QR codes). Electronic identity documents (also called eID documents) are a significant example: they can be used alone without requiring a password-based authentication and are considered as a MFA solution since they attest both a possession (the card itself) and a knowledge factor (the related PIN code). Moreover, such documents can also provide a certain degree of assurance on their owners’ identity, thus significantly simplifying the link between digital and real identities in sensitive contexts. Due to these countless features, eID documents are increasingly involved not only in authentication, but also in the so-called “enrollment” – an online procedure adopted by many banks or service providers to let customers subscribe to a service or sign a contract while having a certain degree of assurance on their real identities [3]. The increasing adoption of these protocols and the sensitive contexts in which they are used require the development of automated security analysis techniques to increase the trust of all stakeholders, from end-users to system administrators while striking a good trade-off among security, privacy, usability, compliance, and efficiency.

The research work to be conducted in the thesis aims to make significant contributions to developing methodologies, techniques and tools to both secure-by-design approaches and the security assessment of Identity Management solutions, including:
- The design of enrollment and authentication protocols relying on eID documents and/or open standards (e.g., OIDC and FIDO2) with particular attention to regulatory issues e.g., PSD2 and Know Your Customer best practices in the financial sectors.
- The automated security analysis of authentication and authorization protocols by exploiting state-of-the-art formal techniques to detect violations of security properties along the lines of e.g., the approach described in [4].
- The automated assessment of risks to which authentication and authorization protocols are exposed by extending existing tools such as MuFASA [5].

References:
[1] National Institute of Standards and Technology. Digital Identity Guidelines (Special Publication 800-63). Available at: https://pages.nist.gov/800-63-3/

[2] World Economic Forum. Passwordless Authentication: The next breakthrough in secure digital transformation. Available at: http://www3.weforum.org/docs/WEF_Passwordless_Authentication.pdf

[3] Silvio Ranise, Giada Sciarretta and Alessandro Tomasi. Enroll, and authentication will follow: eID-based enrollment for a customized, secure, and frictionless authentication experience. In: 12th International Symposium on Foundations & Practice of Security (FPS 2019). DOI: 10.1007/978-3-030-45371-8_10

[4] Giada Sciarretta, Roberto Carbone, Silvio Ranise, Luca Viganò. Design, Formal Specification and Analysis of Multi-Factor Authentication Solutions with a Single Sign-On Experience. In: Principles of Security and Trust (POST 2018). DOI: 10.1007/978-3-319-89722-6_8

[5] Federico Sinigaglia, Roberto Carbone, Gabriele Costa, Silvio Ranise. MuFASA: A Tool for High-level Specification and Analysis of Multi-factor Authentication Protocols. In: Emerging Technologies for Authorization and Authentication (ETAA@ESORICS 2019). DOI: 10.1007/978-3-030-39749-4_9

Space and avionics systems are reaching an unprecedented degree of complexity. The process of safety analysis attempts to characterize the likelihood of faults and failures, and to assess the effectiveness of the adopted mitigation measures. Unfortunately, traditional techniques are becoming ineffective, unable to deal with large-scale systems. This thesis will investigate novel methods for safety analysis, based on the adoption of formal models of system and software (nominal and faulty) behaviors. Particularly interesting are the analysis of timing aspects in the propagation of multiple faults to failures and errors, the ability to explain the causality of propagation, and the definition of techniques for on-the-fly fault detection, isolation and recovery policies.

The proliferation of digital and human-centric systems, spurred by the introduction of 5G communications, Internet of Things (IoT) and Cloud Computing technologies, is posing unprecedented challenges in terms of cybersecurity, with plenty of software and hardware vulnerabilities that attackers are eager to find and exploit. In this context, Artificial Intelligence (AI) is offering powerful instruments (e.g. Deep Learning) to effectively detect attack patterns and discover threats.
With this PhD scholarship, we are seeking candidates that will carry out research and development activities aimed at defining and implementing novel AI-based methods and platforms to secure modern networked systems.
We are therefore encouraging applications from candidates with solid knowledge of the networking, security and AI basics, skilled in programming, eager to work within a vibrant team and on cutting-edge technologies. Successful candidates will also have the opportunity to develop relevant expertise on Software Defined Networks (SDN) and programmable data planes (extended Berkeley Packet Filter - eBPF, P4), as well as on cloud-native technologies (microservices and serverless architectures, linux containers, container orchestration).

Current approaches in statistical relational learning are based on undirected graphical models such as Markov Logic Networks. State of the art algorithms for statistical inference cover the Maximum Likelihood (ML) and Maximum a Posteriori (MAP) tasks, but not so much attention has been devoted to Bayesian Inference. Due to the high complexity of the models that can be generated, statistical inference is approximated using sampling methods. Recently, we proposed a study about Bayesian Inference in hybrid graphical models (i.e., models composed of discrete and continuous random variables); the advantage of Bayesian inference is that, it’s a truly statistical inference and it is very robust to overfitting training data. We design a variational method to solve the “exact inference”. However, to perform Bayesian inference, combinatorial problems on the discrete variables must be solved in a more efficient way, and this is still an open problem. The objective of this thesis, is to extend such proposals and to make scalable.

Evidence-based research is showing positive effects of virtual reality interventions on patients with cognitive decline, in terms of both stimulation and relaxation, targeting key outcome variables such as cognition and emotion. Semi-immersive technology is a promising and effective tool to use with older adults affected by mild cognitive decline or dementia, for delivering interventions able to improve cognitive and routine functions, stimulating patients’ brains in rich multisensory settings. However, recent findings show also the need for developing comprehensive guidelines to develop and implement safe, effective VR interventions for people with cognitive decline. The aim of the PhD project is to deploy co-design methods and guidelines for the design and evaluation of semi-immersive VR environments for patients with cognitive decline in collaboration with patients’ associations and communities. The ideal candidate will be strongly motivated in developing design and evaluation skills in the field of virtual reality applications for healthcare.

Conversational agents are key components of artificial intelligence solutions for the delivery of behavior change interventions in healthcare settings. They provide user-friendly, intuitive interfaces based on dialog-based interaction for supporting the promotion of healthy lifestyles or patients’ compliance with treatment in chronic conditions. The aim of the PhD project is to investigate and deploy effective methodologies for the design and testing of conversational agents, based on deep learning techniques, to deliver virtual coaching interventions in the healthcare domain. The ideal candidate will be strongly interested in conducting interdisciplinary work in the area of AI for BIC (Behavioral Intervention Technologies).

Althought 5G has just arrived, the research towards 6G mobile networks has already started. 5G paved the way towards a connected world where multiple verticals (e.g., automotive, industry, and health), each characterised by different performance targets in terms of bitrate, latency, and reliability, can coexist on the same infrastructure. 6G networks will push this paradigm even futher and will require a paradigm shift in the way mobile networks are deployed and operated.

With this fully funded PhD position we are looking for a candidate willing to work on cutting edge research in the field of 6G mobile with a particular focus on data-driven approaches for network management and operation (slicing, edge/fog computing, isolation, resiliency, etc).

The successful candidate has obtained a master's degree with excellent marks in computer science, is proficient in networking and programming, has an affinity for algorithm design and artificial intelligence, and enjoys working in a multi-disciplinary project. In particular, evidence of hands-on experience with software-defined networking for RANs and open-source LTE/5G stacks is considered to be an advantage.

We are looking for candidates willing to develop novel methodologies based on machine learning, deep learning, pattern recognition and artificial intelligence for information extraction, classification, target detection and change detection in radar images.
The candidate will be requested to deal with images acquired from active systems including Synthetic Aperture Radar (SAR) images acquired from Earth Observation satellite missions, and sub-surface radar sounding data from airborne Earth Observation missions and satellite planetary exploration missions. The latter activity is developed in the framework of the Radar for Icy Moons Exploration (RIME) payload on board of European Space Agency (ESA) JUpiter ICy moons Explorer (JUICE) and Sub-surface Radar Souder (SRS) payload under development for the European Space Agency (ESA) Europe's Revolutionary Mission to Venus (EnVision).
Besides the requirements established by the rules of the ICT school, preferential characteristics for candidates for this scholarship are:
• master degree in Electrical Engineering, Communication Engineering, Computer Science, Mathematics or equivalents;
• knowledge in pattern recognition, image/signal processing, statistic/remote sensing/radar.

We are looking for candidates willing to develop novel methodologies based on machine learning, deep learning pattern recognition and artificial intelligence for information extraction, classification, target detection and change detection in multi-/hyper-temporal remote sensing images.
The candidate will be requested to deal with both multi-/hyper-spectral images acquired by passive satellite sensors and Synthetic Aperture Radar (SAR) images acquired from active systems for Earth Observation. The goal is to design novel methods able to use temporal correlation to model landcover behavior, changes and trends for a better understanding of phenomena over time and of climate change.
Besides the requirements established by the rules of the ICT school, preferential characteristics for candidates for this scholarship are:
• master degree in Electrical Engineering, Communication Engineering, Computer Science, Mathematics or equivalents;
• knowledge in pattern recognition, image/signal processing, statistic/remote sensing, passive/active sensors.

This PhD project has the ambition to explore the fusion of multiple modalities and the design of novel cross-modal deep neural network architectures to analyze and generate behaviors and social interactions in multiple settings (e.g. urban scenarios, workplace settings, etc.). Special emphasis will be given to exploit synthetically generated data within GAN architectures. The ideal candidate will be strongly motivated in developing skills in machine learning with a special focus on deep learning, in computer vision and multimodal approaches as well as in human behavior understanding. The project will be supervised by Bruno Lepri (FBK).

The increasing popularity of social media platforms like Twitter and Facebook has led to a rise in the presence of hate and aggressive speech on these platforms. Despite the number of approaches recently proposed in the Natural Language Processing research area for detecting these forms of abusive language, the issue of identifying hate speech at scale is still an unsolved problem. In particular, current hate speech detection systems do not perform well on under-resource languages, are not able to generalise well across different platforms, and fail to integrate contextual information (e.g. network structure, discourse context, links to external media). We are therefore looking for candidates with strong interest in hate speech detection using deep learning techniques that would contribute to the development of novel approaches for robust hate speech detection designed to work in multilingual settings with small or no language-specific training data (zero-shot learning).

We are looking for a candidate willing to embark a challenging activity focused on the development of innovative Microsystems-based Radio Frequency RF-MEMS passive components and networks for next generation of telecommunications, wireless and radio systems and applications, like 5G and the Internet of Things (IoT).
Capitalising on the fully in-house RF-MEMS technology, the candidate will have the opportunity to focus on different stages of prototypes development, from the elaboration of novel RF-MEMS design concepts, to the multi-physical simulation, modelling, fabrication, experimental testing of physical samples and integration.

Embedded systems are a fundamental component of our world. Their ubiquitous adoption and ever-increasing complexity makes the task of verifying their correctness both extremely challenging and extremely important. Techniques based on formal methods and automated theorem proving (particularly in the form of Satisfiability Modulo Theories - SMT) are very appealing in this context, promising to deliver both a higher level of confidence than traditional techniques and a high degree of automation. The objective of this PhD research is that of advancing the state of the art in the application of SMT-based formal methods to parameterized systems -- systems consisting of an unbounded number of components/processes -- which naturally arise in many safety-critical domains. The candidate will work on both theoretical aspects of the problem, as well as its practical applications in relevant case studies, drawn from the domains of railways, avionics and aerospace.

The need to translate the audio from one language into a text in a target language has dramatically increased in the last few years with the growth of audiovisual content freely available on the Web. Current speech translation (ST) systems need to be able to serve different applications working in various scenarios and to satisfy several factors coming from the market (e.g. specific length of the output, adaptation to different domains, real-time processing) or present in the audio (e.g. background noise or strong accent of the speaker). The objective of this PhD is to advance the state of the art in speech translation to make ST flexible and robust to these and other factors. Candidates should have a strong curiosity to solve problems in natural language processing and have a background in deep learning and maths, as well as excellent programming skills in Python. They will work both on theoretical aspects of the problem, and on their practical application in relevant case studies driven by ongoing projects where the MT unit is involved. Applicants are invited to contact us ([email protected] and [email protected]) in advance for preliminary interviews.

The Internet of Things paradigm is driving innovation in several application domains, such as digital industries, healthcare, smart cities, energy, retail, agriculture, etc. IoT architectures have been evolving from an initial centralized approach towards more distributed and edge-centric ones, where everyday-more-autonomous IoT devices are able to interact in a decentralized and effective way. However, the huge variability of requirements that an IoT platform needs to cope with and fulfill to serve such diverse application domains and scenarios suggests that a holistic architectural approach cannot be the solution. Indeed, depending on the specific application context, different challenges emerge, while often conflicting requirements need to be addressed meeting the best trade-offs between energy management and conservation, cognitive capabilities deployed on autonomous devices, latency and bandwidth over low-power communication protocols, timeliness of data exchange, reactivity and reliability in wide area networks.
The objective of this PhD is to study, analyze and identify novel energy-aware architectural approaches capable to balance the use of available energy resources within constrained IoT devices with embedded computing capabilities and possible strategies to overcome the existing technological limitations by using blended strategies, such as combination of energy harvesting technologies, computation offloading from the edge to the cloud, adaptive data exchange paradigms, and distributed artificial intelligence along the cloud-to-thing continuum.

The Phd project can evolve along either one of the two research lines described below.

Research line 1: Personal agents for mindfulness-based interventions in healthcare. In the last decade there has been a growing interest for mindfulness interventions delivered by means of mobile applications and personal assistants to support self-care of patients, including those coping with chronic conditions. Despite the fact that the validity of mindful practices has been proved repeatedly by previous research, the design of effective behavioral intervention technologies for mindfulness coaching and practice remains a challenge. The aim of the PhD project is to investigate key features of smart coaching solutions for mindful-based interventions that are engaging to use by patients and produce effective outcomes from a clinical perspective. The ideal candidate will be strongly motivated in developing design skills in the field of behavioral intervention technologies and conversational agents for applications in healthcare.

Research line 2: Computational creativity is a sub-field of artificial intelligence concerned with the development of programs that can produce creative output;
in particular, many of these programs deal with linguistic creativity.
Recent years have witnessed a growing interest in computational linguistic creativity, a research field at the boundary between many disciplines including
natural language processing, linguistics, psychology, and of course cognitive sciences. Even though the state-of-the-art in this field is advancing a bit in the last years, real-world applications of computational linguistic creativity are still uncommon. Particular focus of the PhD will be put in computational treatment of figurative, emotional, witty language considering also in multimodal (e.g. visual) context.

Responsible for the project: Silvia Gabrielli / [email protected] (for line 1) and Carlo Strapparava [email protected] (for line 2)

With the increasing availability of powerful mobile and ubiquitous computing technologies, motivational and behavior change technologies, traditionally applied in the health science domain, have become commonplace and have expanded to cover several other domains. In particular, game-based motivational systems (i.e., serious games, games with a purpose, gamification) are emerging as an effective tool to induce a positive change in human behaviors in relation to environmental challenges, social engagement, safety, productivity and learning. Games introduce goals, interaction, feedback, problem solving, competition, narrative, and fun learning environments, motivational affordances that can increase end-user engagement and induce a voluntary behavior change. Although several research studies proved the validity of game-based motivational systems, their effectiveness has often been context specific and varied among individuals. A promising and ever-growing research field concerns the investigation of user-centered, personalized and adaptive gamified dynamics, tailored to specific users and contexts. In the context of games, adaptivity describes the automatic adaptation of game elements, i.e., of content, user interfaces, game mechanics, game difficulty, etc., to customize or personalize the interactive experience. Personalization and adaptivity can promote motivated usage, increased user acceptance, and user identification in serious games.

The aim of the PhD project is to investigate key features of adaptive personalized motivational game-based systems addressing one or more of the following research challenges: analyse the relationship between the mechanics and their effects on different individuals to react accordingly; investigate adaptive game dynamics in multi-player settings, finding the right balance between individual (micro) level and community (macro) level personalization; design of meaningful adaptive gamified reinforcement strategy to sustain players’ long-term motivation (e.g., churn prediction and generation of dynamic personalized re-engagement incentives).

The PhD project focuses on the cognitive basis of computational thinking for the design of digital technologies in the field of Internet-Of-Things aimed at improving learning and education. In particular, the research explores digital tools that can be manipulated and adapted by teachers and children themselves (the so-called "End-User Programming").

Research activity conducted during the Ph.D. will focus on the development of computational tools aimed at providing insight into factors influencing the spatio-temporal spread of infectious disease in humans and reliable estimates of the expected impact of public health interventions. Envisioned approaches range from the study of mechanistic models mimicking the transmission processes underlying outbreak, epidemic, and pandemic situations to the use of statistical inference applied to epidemiological data.

A clinical case is a statement of a clinical practice, presenting the reason for a clinical visit, the description of physical exams, and the assessment of the patient’s situation. Clinical cases (e.g. discharge summaries, clinical cases published in journals, and clinical cases from medical training resources) provide a very valuable source of information for data-driven technologies aiming at predicting clinical outcomes and patient behaviors.

This three-year PhD offers a unique context of a collaboration between FBK, specifically the NLP group, and a Swiss multinational healthcare company, worldwide leader in biomedical research.

The research will focus on a number of application oriented tasks, including automatic recognition of clinical entities (e.g. pathologies, symptoms, procedures, and body parts, according to standard clinical taxonomies such as ICD-9, ICD-10 and SNOMED-CT); detection of temporal information (i.e. events, time expressions and temporal relations, according to the THYME TimeML standard), and factuality information (e.g. event factuality values, assessment of the effect of negation, uncertainty and hedge expressions).

Italian will be the major language of clinical cases, although technologies will be experimented on other languages.

The goal of the PhD is both to advance the state of the art for clinical case analysis for the Italian language, and to deliver prototype applications, which can be further made operative in real settings (e.g. hospitals).
The candidate will have the unique opportunity to explore different domains (Natural Language Processing, Machine Learning, Health & Well-Being) being directly coached by very experienced teammates.

The involved PhD will work in an international environment, collaborating with a healthcare company, with worldwide presence. The candidate will work both at FBK (Trento) and at the abovementioned company’s premises (both in Italy and abroad).

The assurance of safety requirements is fundamental in many safety-critical domains. In particular, in the automotive domain, the increased automation and connectivity of embedded systems demand for more rigorous analysis of their functional safety. The advances in sensor and actuator technologies enable new solutions to increase the safety and reliability of systems. At the same time, their complexity makes the assurance of such properties more and more challenging.

This thesis will concentrate on developing new formal methods for the design of safety solutions in automated systems. It will investigate advances with respect to the state of the art on formal methods about the specification and verification of safety contracts, the diagnosis and prognosis of their failure, and the related diagnosability problem. The validation of the methods will be performed with case studies developed in automotive domain and based on real-world solutions developed by Bosch.

The correct specification of requirements is a cornerstone in the design of safety-critical systems and in the assurance of safety requirements. Errors and ambiguities in the requirements are considered a fundamental problem by many industries across application domains. Rigorous techniques based on formal methods have been shown to be promising to address this problem. However, these techniques are difficult to take up at the industrial level. There is still a large gap between the information requirements and their formal counterpart. Each domain has specific ways to specify the requirements that depend also on the kind of system or component under design. Moreover, the semantic interpretation is also very dependent on the domain, which typically has some assumptions and background knowledge that are very expensive to formalize.

This thesis will focus on the development of a new open-source framework based on formal methods for the formalization and validation of domain-specific requirements. The work will support the creation of domain-specific control natural languages for requirements. Moreover, it will provide means to specify the domain-specific background knowledge and to exploit for tailored analysis engine. The framework will be validated with case studies provided by Intecs, which with its experience in diversified domains (railways, aerospace, IoT, etc) will be a suitable environment for this purpose.

Digital technologies play an ever increasing role in all aspects of human society; this induces a wide range of changes, collectively referred to as Digital Transformation, that, far from being only technological, also cover cultural, organizational, social, managerial aspects of our life.
Artificial Intelligence is a key technology for digital transformation, thanks to its capability to extract information and knowledge from data; this requires the capability to open, analyze and exploit all data available on a given phenomenon, data that are often highly heterogeneous, scattered, and coming from different sources (e.g. open, sensor, free, closed, linked data). This thesis will concentrate on developing a data-driven computational framework, based on AI approaches, able to perform data analysis and prediction in the setting just described. The framework will be developed in the scope of the Digital Hub, a digital platform jointly developed by Dedagroup and Fondazione Bruno Kessler to address digital transformation in different application domains, including Public Administration, Digital Finance, Digital Industry. The validation of the framework will be performed addressing problems in these application domains, by exploiting the data sets and services integrated in the Digital Hub.

This project will focus on data science and deep learning to quantify and model observations made within experiments of particle physics (ATLAS at the LHC, future colliders) and in space (HEPD-01/02, Aladino). The project will develop novel analytics tools for event building software, including feature tagging, consequent event selection and particle identification.

The candidate will perform research on highly correlated and interdependent networks, providing a model for many empirical complex systems, natural and artificial. The underlying framework is based on statistical physics at and out of equilibrium, and nonlinear dynamics. The candidate will develop the theoretical and computational ground for the analysis of collective phenomena emerging in such systems, with application to biological, social and urban networks.

The PhD grant has the goal of developing methodologies, based on statistical physics, for modeling complex systems such as socio-technical, urban, or economic systems. The developed multilayers models will be adopted to investigate and explain the complex characteristics of the observed socio-economic phenomena with the goal of identifying a minimal set of microscopic mechanisms able to reproduce collective phenomena on meso- or macroscopic scales.

Conversational agents are supposed to help users to accomplish common tasks, like booking a restaurant, buying a product, give commands while driving a car, or suggesting correct life styles.
In the last years several methods based on deep learning architectures have been developed to build conversational agents that learn their behavior directly from data (i.e. dialogues). Although this research line is very promising, several challenges are still open, including (i) how to incorporate domain knowledge into end-to-end architectures; (ii) how to reproduce human-like conversational behaviors, particularly those related to collaboration, and (iii) how to address the scarcity of annotated dialogues for many application domains and languages.
This PhD grant aims at advancing the state-of-the-art in selected topics in data-driven conversational agents, building on top of current research and projects carried on at the NLP group at FBK.

Process mining is a family of methods that aim at analyzing business processes based on their observed behavior recorded in event logs. Sometimes, however, together with data related to the observed behaviour, additional knowledge about the process execution is also available (e.g., knowledge provided by domain experts, knowledge related to rules that have to be observed). This knowledge can be leveraged in order to improve the accuracy of the results of state-of-the-art process mining techniques, e.g., to discover more accurate process models or to provide more accurate prediction.
Within this PhD program, the candidate will work towards combining the information gathered from recorded event logs and additional knowledge in the context of different process mining scenarios, such as process discovery and predictive process monitoring. Achieving this task requires dealing with several challenges as, for example, the capability to envision new algorithms able to combine data and knowledge and the capability to implement scalable solutions.
The work will put together theoretical and methodological aspects, ranging from the capability to identify the correct representation for a complex problem up to the development of process mining algorithms and tools.

The ability of modeling, understanding and predicting human behaviors and social interactions is fundamental for computational social science and has a range of relevant applications for individuals, companies, and societies at large. In this project, the goal is merging approaches from network science and machine learning and using data on mobility routines (e.g. GPS and other mobile phone data), face-to-face interactions and purchase behaviors (e.g. credit card transactions) in order to develop methods for quantify daily habits, individual dispositions and traits, and behavioral changes. A special attention will be given to the changes on daily human behaviors due to the emergence and spread of the Covid-19 pandemic. The Ph.D. project will be conducted within the FBK MobS research unit but with collaborations with several international groups.

SPAD and SiPM mixed-signal interfaces for LiDAR pose several challenges in terms of low noise, low power and speed. Clever architectures and optimized designs are necessary to overcome the current limitations: this PhD position aims at investigating the next-generation interfaces for SPAD and SiPM, focused on the distance ranging application but also applicable to the readout frontends for particle physics and medical instrumentation. The activity spans from the modeling aspect, to the readout IC design, simulation and characterization, to the system integration, requiring electronics design skills but also knowledge of optics and discrete electronics.

Automatic semantic segmentation of images and 3D data is becoming a very prominent research field with many interesting and reliable solutions already available. To this end, machine learning approaches are fundamental to automatize the processing and to couple geometric data with semantic labels. The aim of the PhD is to advance the state-of-the-art in semantic photogrammetry (including SLAM and MVS) applied to industrial, heritage and territorial applications where real-time is requested. Within the real-time image processing we want to embed classification solutions in order to offer semantic 3D data useful for e.g. inspection, quality control, monitoring, etc.

Automated Planning is the problem of synthesizing courses of actions guaranteed to achieve a desired objective, given a formal model of the system being controlled. A class of problems particularly interesting for applications is temporal planning (also called planning and scheduling) where the discrete decisions of "what to do" are coupled with the problem of scheduling (deciding "when to do"). Unfortunately, planning and scheduling techniques suffer from scalability issues and are often unable to cope with the complexity of real-word scenarios, despite the plethora of approaches available in the literature.
Recently, efforts such as Deepmind AlphaGO and OpenAI Five hit the headlines, with groundbreaking advancements in the field of reinforcement learning. These techniques are able to automatically learn policies to decide what to do in order to achieve a desired goal. However, they offer no formal guarantee and are not model-based. The research objective of this PhD scholarship is to investigate techniques that combine the formal guarantees offered by automated planning and scheduling with the performance and self-improving capabilities offered by recent advances in deep reinforcement learning to construct self-adaptive planners that can improve over time their performance on specific application scenarios.

Silicon Photomultipliers (SiPMs) are silicon photodetectors with internal signal amplification obtained by means of the impact ionization mechanism, operating in the so-called Geiger mode and achieving a gain in the order of a few millions per a single, photo-generated carrier. They are extremely fast and sensitive detector, having demonstrated the capability of detecting light down to the single photon level and with a Single Photon Time Resolution (SPTR) of 20 ps FWHM.
Thanks to these exceptional characteristics, SiPMs are currently of great interest in several scientific and industrial applications. They are being considered for the detection of faint light signals in the large majority of big scientific experiments of the next generation, ranging from the HL-LHC upgrade to future Dark Matter and Neutrino experiments. For these experiments, SiPM technology needs to be developed further, to obtain, among other features, enhanced radiation hardness, effective operation at cryogenic temperatures and sensitivity to extremely low-wavelength photons down to 128 nm, for direct detection of LAr scintillation light. By overcoming such difficult challenges, SiPMs, will constitute an enabling technology for scientific research over the next decade.
In this context, Fondazione Bruno Kessler (FBK, Trento) is at the forefront of developments. FBK is currently collaborating with CERN and several other research institutions for the development of rad-hard SiPMs, which are required to withstand radiation doses in excess of 1e14 neq/cm2. Considering cryogenic operation of SiPMs, R&D carried out by FBK for experiments such as DarkSide-20k (http://darkside.lngs.infn.it/), nEXO (https://nexo.llnl.gov/) and DUNE (https://www.dunescience.org/) has already achieved outstanding results, such as almost noiseless detector operation at 87 K (LAr temperature). On the other hand, research is still ongoing to improve detector characteristics cold temperatures, such as reducing afterpulsing, output capacitance and increasing photon detection efficiency (PDE).
The PhD candidate will work in the IRIS research unit of FBK (https://iris.fbk.eu/), which is composed of approximately 20 people, mostly researchers and PhD students, working on SiPM, SPAD and CMOS image sensors development. Within IRIS, his/her research activity will be focused on the study and development of SiPMs built in custom technology in FBK clean-room. The PhD candidate will be inserted in a team of researchers developing word-class sensor technologies, with the purpose of becoming proficient in the field of instrumentation and, in particular, of silicon sensors and readout techniques. He/She will learn how to characterize and optimize the components of a photon or radiation detection system, taking into account the requirements of different scientific and industrial applications.
The PhD candidate will dedicate his/her research to one or a few scientific projects, selected together with his/her supervisor among the ones carried out by FBK and characterized by uniform technical challenges. At the beginning of the three-year studies, he/she will be focused on understanding the SiPM working principle and its main characteristics and on becoming expert in some of the different, advanced characterization setups and techniques employed at FBK. They include: semiconductor analyzers, high-speed digitizing oscilloscopes, ultra-fast laser sources, thermostatic chambers, a cryostat, instrument programming in Labview, digital data processing, data analysis, development of custom, front-end electronics, scintillating materials and radioactive gamma-ray sources. An important part of his/her activities will be constituted by the advanced characterization of different SiPM process and layout splits, which are usually included in R&D runs fabricated at FBK. Such activity will be aimed at understanding the features of the device, studying the different phenomena involved, evaluating the effectiveness of the technical solution implemented and proposing further developments of the sensor. After acquiring sufficient experience, towards the end of the PhD studies, it will be also possible for him/her to contribute to the design of new SiPM sensors.
At the end of the PhD studies, the PhD candidate is expected to become proficient in the field of sensors, instrumentation and characterization setups. He/She will be able to start a career either in the scientific field, for example joining a large scientific collaboration, in which his/her competences are usually highly valued.
For any further question or curiosity, please contact Dr. Alberto Gola, [email protected].

Silicon Photomultipliers (SiPMs) are silicon photodetectors with internal signal amplification obtained by means of the impact ionization mechanism, operating in the so-called Geiger mode and achieving a gain in the order of a few millions per a single, photo-generated carrier. They are extremely fast and sensitive detector, having demonstrated the capability of detecting light down to the single photon level and with a Single Photon Time Resolution (SPTR) of 20 ps FWHM.
Thanks to these exceptional characteristics, SiPMs are currently of great interest in several scientific and industrial applications. They are being considered for the detection of faint light signals in the large majority of big scientific experiments of the next generation, ranging from the HL-LHC upgrade to future Dark Matter and Neutrino experiments. On the other hand, SiPMs have also very important industrial applications, with great potential to improve health and several aspects of people’s lives. The most traditional industrial application is the use of SiPMs to build next-generation, Time-of-Flight Positron Emission Tomography machines (ToF-PET), which are used in neurology, cardiology, oncology, drug development, and in a number of other medical imaging use cases. A second, extremely interesting SiPM application is LIDAR for automotive. Indeed, SiPMs are considered one of the most promising technology solutions to enable LIDAR sensors that will provide 3D vision to next-generation, autonomous driving systems (level 4 and above).
In this context, Fondazione Bruno Kessler (FBK, Trento) is at the forefront of developments. FBK is currently collaborating with CERN and several other research institutions for the development of rad-hard SiPMs, which are required to withstand radiation doses in excess of 1e14 neq/cm², and of cryogenic SiPMs, for experiments such as: DarkSide-20k (http://darkside.lngs.infn.it/), nEXO (https://nexo.llnl.gov/) and DUNE (https://www.dunescience.org/). Collaboration with companies is also very active, with ongoing projects related to development of next-generation SiPMs for both ToF-PET and LIDAR.
The PhD candidate will work in the IRIS research unit of FBK (https://iris.fbk.eu/), which is composed of approximately 20 people, mostly researchers and PhD students, working on SiPM, SPAD and CMOS image sensors development. Within IRIS, his/her research activity will be focused on the study and development of SiPMs built in custom technology in FBK clean-room, with special attention dedicated to the interaction between sensor and front end, to advanced 3D integrated and BSI SiPM structures. In this context, the partnership between FBK, which develops SiPMs, and Politecnico di Torino and INFN, sezione di Torino, which are expert in CMOS ASIC design, is strategic. The PhD candidate will be inserted in a team of researchers developing word-class sensor technologies, with the purpose of becoming proficient in the field of instrumentation and, in particular, of silicon sensors and readout techniques. He/She will learn how to characterize and optimize the components of a photon or radiation detection system, taking into account the requirements of different scientific and industrial applications.
The PhD candidate will dedicate his/her research to one or a few scientific projects, selected together with his/her supervisor among the ones carried out by FBK and characterized by uniform technical challenges. At the beginning of the three-year studies, he/she will be focused on understanding the SiPM working principle and its main characteristics and on becoming expert in some of the different, advanced characterization setups and techniques employed at FBK. They include: semiconductor analyzers, high-speed digitizing oscilloscopes, ultra-fast laser sources, thermostatic chambers, a cryostat, instrument programming in Labview, digital data processing, data analysis, development of custom, front-end electronics, scintillating materials and radioactive gamma-ray sources. An important part of his/her activities will be constituted by the advanced characterization of different SiPM process and layout splits, which are usually included in R&D runs fabricated at FBK. Such activity will be aimed at understanding the features of the device, studying the different phenomena involved, evaluating the effectiveness of the technical solution implemented and proposing further developments of the sensor. After acquiring sufficient experience, towards the end of the PhD studies, it will be also possible for him/her to contribute to the design of new SiPM sensors.
At the end of the PhD studies, the PhD candidate is expected to become proficient in the field of sensors, instrumentation and characterization setups. He/She will be able to start a career either in the scientific field, for example joining a large scientific collaboration, in which his/her competences are usually highly valued.
For any further question or curiosity, please contact Dr. Alberto Gola, [email protected].