PhD course in
Computer Science and Artificial Intelligence

Call 2024

(Open until June 20, 2024)

The Department of Mathematics, Computer Science and Physics of the University of Udine hosts the PhD course in Computer Science and Artificial Intelligence in agreement with Fondazione Bruno Kessler. The course continues an outstanding tradition in computer science teaching and research at the University of Udine, and ideally links up with the best science education courses in Italy at master level, as stated by the 2020/21 official ranking made by CENSIS. This tradition is further enriched by the dynamic, project-oriented knowledge production running at Fondazione Bruno Kessler, generating an ideal environment where top students can meet excellence in both theoretical as well as applied research fields.

The course is active since the XXXVII cycle (2021/22), and originates from the splitting of the PhD course in Computer Science, Mathematics and Physics. It resumes the tradition of the previous PhD course in Computer Science, active across thirty years since the first national cycle (1983/84) until the XXX cycle (2012/13).

Training goals

The PhD course in Computer Science and Artificial Intelligence will graduate students with top skills, in topics that are listed below with links to the involved scientists belonging to the PhD Board, also in the context of a multi-disciplinary research plan:

  • Acoustic scene analysis and Machine listening
  • Algorithms
  • Artificial intelligence in agrifood
  • Automatic planning and scheduling
  • Autonomous systems
  • Blockchain and Digital ledger technologies
  • Computational biology and Bioinformatics
  • Computational intelligence and Optimization
  • Computer vision
  • Crowdsourcing
  • Cyber-security
  • Data science and Big data analytics
  • Digital humanities
  • 3D digitalization with Artificial intelligence
  • Distributed systems: models and applications
  • Formal methods and Automatic verification
  • Human-Computer interaction, Auditory-tactile interfaces
  • Knowledge representation and Automatic reasoning
  • Information retrieval
  • Internet of things: platforms and technologies
  • Logics in computer science
  • Machine learning and Deep learning
  • Medical informatics, Tele-medicine and e-Health
  • Methodologies, languages and techniques for problem solving in artificial intelligence
  • Natural language processing
  • Predictive monitoring, diagnostics and maintenance
  • Social systems and Recommendation systems
  • Software engineering
  • Virtual reality, Serious games

External scholarships
(Bruno Kessler Foundation, FBK)

1) 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. 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, that played a major role during the ongoing pandemics, and the Center for Digital Industry, a leading center in model-based design.

Supervisor: Alessandro Cimatti

2) Methodologies for parametric systems testing

Modern software systems are often characterized by a high degree of variability in the possible functional configurations and release architectures. This causes a dramatic increase in complexity, and new methods and tools are required to design reliable parametric software systems.

The goal of this PhD thesis is to explore new approaches to modeling, testing, verification and validation of parametric systems, involving the joint use of model-based and AI based techniques such as formal methods, machine learning and optimization.

Supervisors: Angelo Susi, Alessandro Cimatti

3) Epistemic Runtime Verification

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.

Supervisor: Alessandro Cimatti

4) 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.

Supervisor: Alessandro Cimatti

5) 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.

Supervisor: Alessandro Cimatti

6) Meta-learning for advanced 3D representations

Learning-based algorithms for 3D object reconstruction, recognition, classification 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 hamper the employment of learning-based 3D processing pipelines in daily industrial practices or for large-scale real-world applications. with data coming from urban and environmental contexts.

The goal of the PhD position is to conduct research on novel and efficient algorithms based on zero-shot learning and diffusion models applied to 3D reconstruction and classification tasks. The aims include:

  1. to explore state-of-the-art methods for learning-based 3D representation, including reconstruction, generation, segmentation and classification
  2. to conduct disruptive research on efficient methods for 3D reconstruction and understanding of complex surfaces, including reflective and transparent surfaces
  3. analyse, test and demonstrate replicability in multiple real-life scenarios, including industrial objects.

Supervisor: Fabio Remondino

7) Multi-modal learning-based SLAM

Mobile mapping solutions, being hand-held or vehicle-based, are becoming more and more widespread in surveying, robotics and autonomous vehicles. These solutions, based on Simultaneous Localization and Mapping (SLAM) techniques, incorporate an array of sensors including visual, LiDAR, inertial and GNSS among others. However, mainstream algorithms lack the simultaneous utilization of the rich sensors suite (i.e. 3 or more of them). Furthermore, their applicability across diverse conditions (forest, urban, indoor, unstructured, mixed, etc.) remains a challenge, particularly for deep-learning-based solutions. This PhD project aims to:

  1. explore state-of-the-art methods for all components of SLAM (e.g., visual- or LiDAR-based odometry, loop closure detection and matching, multi-sensor data adjustment), including both handcrafted and learning-based approaches;
  2. develop a robust and universal multi-modal SLAM framework for jointly processing the data acquired by all on-board sensors, preferably utilizing bundle adjustment concepts, with the aim of challenging state-of-the-art single-data alternatives in terms of resulting accuracy;
  3. carry out experimental verifications with robotics and hand-held solutions of the quality of the 3D data acquired by the developed system in varying and challenging conditions.

Supervisor: Fabio Remondino

8) AI-based Models and Tools for Next-Generation Serious Games

The PhD activity will offer an opportunity for motivated candidates to embark on a PhD journey delving into the realm of next-generation serious games (SGs) infused with Artificial Intelligence (AI). Focused on revolutionizing game development, particularly in educational and sustainability domains, the main aim is to explore the transformative potential of AI advancements. By exploiting AI-driven capabilities, SGs can dynamically adapt to individual player profiles, fostering deeper engagement and immersion. Central to this initiative is the concept of human-AI collaboration, where AI systems seamlessly integrate with human players to provide tailored and challenging experiences. The PhD grant sets out to achieve key objectives including AI-driven content generation, adaptive gameplay mechanisms, personalized game experiences, and collaborative human-AI interaction. Moreover, it addresses ethical and societal implications, ensuring responsible and inclusive game design. Through interdisciplinary research and European collaborations, the PhD activity aims to advance AI-driven serious game development, creating a new generation of impactful and immersive SGs.

Supervisors: Antonio Bucchiarone, Kevin Roitero

9) Pareto-based optimization methods to support one-click deployments of EdgeAI application flows

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.

Supervisor: Massimo Vecchio

External scholarships (FSE Call)

1) Study and development of Artificial Intelligence techniques for optimising the water use and energy consumption of industrial plant

Several companies in the country have plants with advanced production facilities that require a large use of energy, mainly electricity, for their operation, and in many cases are characterised by high water consumption. The challenge for the “better 203″0, i.e. the climate neutrality of production activities, confronts companies, especially large ones, with the need to articulate a programme based on continuous improvement in terms of energy efficiency and reduction of water consumption in production processes. The main objective of the proposed research activity is to study and develop Artificial Intelligence techniques cable of improving the energy efficiency of industrial production plants. The proposal is based on the idea that reducing the energy consumption of production processes does not necessarily imply lower production, but rather an improvement in the productivity and effectiveness of plants, allowing the same amount of product to be produced with fewer resources.

Supervisor: Gian Luca Foresti

2) Multisensory interactions and auditory/tactile interfaces for rendering digital experiences beyond vision

An increasing number of regional young companies and startups are participating in the development and marketing of virtual reality scenarios for the public, such as those that will see the light at the Digital Experience Centre of the Maritime Museum in Trieste. Completing visual feedback with 3D audio and somatosensory elements is key for realizing virtual experiences hosted by future museums and entertainment spaces. The PhD candidate will research concepts, prototype software and adapt existing hardware around interaction components at the intersection between sonic and haptic interaction design, with the goal of enabling novel multisensory virtual objects and scenes having also corporate interest.

Supervisor: Federico Fontana

3) XAI-FVG Explainability of Weather Forecasting in FVG

Following an experience gained from a multi-year collaboration with researchers from the ARPAFVG (Osmer), in particular for neural network models for the prediction of extreme events (lightning, hail), we wish to study the possible application of automatic symbolic AI techniques , both programmed upstream with expert knowledge and automatically and dynamically extracted from the data, to explain the reasons for the predictions (Explainable AI, or XAI) applicable both to the aforementioned extreme events and to daily forecasts on a regional basis. These explanations may also be useful for modifying the sub-symbolic models used so far as a result of rapid climate variations.

Supervisor: Agostino Dovier

4) The Role of New Technologies in the Green Deal: More Efficient Models for Artificial Intelligence and Deep Learning

Artificial Intelligence and Deep Learning systems are crucial for today’s businesses, but their complexity requires a lot of energy. This project aims to develop lighter and more efficient models, without sacrificing performance and accuracy, through optimization techniques, data compression, quantization and intelligent resource management. The primary objective is to encourage the mitigation of environmental impact, offering direct support to the achievement of the Green Deal objectives promoted by the FVG region.

Supervisor: Giuseppe Serra

5) Towards AI Solutions for CSI-based Wireless Sensing and Positioning to Support Pervasive Home and Health Care

The research will explore the possibility of exploiting Channel State Information (CSI) from wireless sources to perform sensing and positioning in complex and critical scenarios, like, for instance, home and health care. The goal is to collect meaningful information, like respiratory rate, heartbeat, people gestures, movements, and location, using minimal, non-dedicated devices only, e.g., smartwatch, smartphone, and access points, to be used to develop a general framework that supports assisted living. Concrete applications range from fall  to abnnormal behaviour detection, from apnea recognition to remote monitoring of the recovery of stroke-affected patients. Despite the potential of CSI, its effectiveness is currently limited by degradation effects that significantly alter signal patterns over time. Thus, to a large extent, the research will focus on studying and developing time- and space- invariant Machine and Deep Learning models to mitigate such problems. The research will be carried out in collaboration with some local hospital medical units.

Supervisor: Andrea Brunello, Angelo Montanari, Nicola Saccomanno

6) Machine Learning methods for Disability Identification in Elecronic Health Records

Aim of the present project is to study methods and techniques to identify and represent disability and frailty signs in electronic health records, by means of machine learning (ML). ML could be applied in two directions: (i) to detect signals in already available text and data or (ii) to support healthcare professionals in coding frailty and disability conditions by means of available classifications such as ICF. The Friuli Venezia Giulia Region has a regressive population structure with a relatively high elderly component, more at risk of frailty. Furthermore, for some disability conditions (in particular motor and deafness) the Region is among those with the highest prevalence. The use of machine learning methods can on the one hand help operators to adequately codify the conditions (for example, also but not only for certification purposes), on the other to recognize signs, for example in the ESF, which can allow provide the necessary supports to people with disabilities, always respecting privacy.

Supervisor: Vincenzo Della Mea

7) Diagnosis of dysphonia and laryngeal pathologies using advanced numerical models of phonation and AI techniques

Aim of the present project is to study methods and techniques to identify and represent disability and frailty signs in electronic health records, by means of machine learning (ML). ML could be applied in two directions: (i) to detect signals in already available text and data or (ii) to support healthcare professionals in coding frailty and disability conditions by means of available classifications such as ICF. FVG has a regressive population structure with a relatively high elderly component, more at risk of frailty. Furthermore, for some disability conditions (in particular motor and deafness) the Region is among those with the highest prevalence. The use of ML methods can on the one hand help operators to adequately codify the conditions (for example, also but not only for certification purposes), on the other to recognize signs, for example in the ESF, which can allow provide the necessary supports to people with disabilities, always respecting privacy.

Supervisor: Carlo Drioli

8) Immersive therapeutics for chronic pain management

The issue of management of chronic pain conditions is increasingly relevant in the region, also due to an increasing ageing of the population. Chronic pain is a major research topic in Virtual Reality (VR), and some applications have been successful in clinical contexts. The current challenge for research is to empower patients with the capability of using VR at home as an immersive therapeutic tool for chronic pain management. Moreover, the possibility of applying Augmented and Mixed reality in addition to VR for chronic pain is still unexplored. The PhD candidate will work at designing, implementing, and evaluating (also in collaboration with expert doctors from regional hospitals) an immersive therapeutics system for home use. Central goals of the system will be a very high level of usability, to allow its use by anyone, including older adults, and a compelling user experience, in order to keep the user engaged over long periods in such a way that the analgesic effect will not wear off over time.

Supervisor: Luca Chittaro

9) SistAnimalID – Animal Recognition System

The development of an animal recognition system based on artificial intelligence can promote innovation and sustainability in the regional livestock and fisheries sector. It would allow for more efficient monitoring of livestock and fish species, improving animal welfare and production efficiency. It would represent a frontier application of intelligent technologies in key sectors of the territory.

Supervisor: Niki Martinel

10) BioSubAcque – Underwater Image Analysis for Environmental Monitoring

Advanced analysis of underwater images via deep learning can support the monitoring of the marine and coastal ecosystems of the FVG. It would promote environmental protection, the sustainable management of fish resources and the development of innovative solutions for the regional Blue Economy. It would contribute to efforts to transition towards a circular blue economy.

Supervisor: Niki Martinel

11) TrustVision – Efficient and Reliable Artificial Vision for Industry 4.0

The development of efficient and certified artificial vision models can unlock new sustainable industrial applications. These intelligent systems can automate production processes, inspect plants and optimize logistics, promoting innovation in local manufacturing companies. Energy waste would be reduced and the circular economy would be increased thanks to predictive monitoring.

Supervisor: Niki Martinel

External scholarships (PNRR call 118)

1) Computer vision for tracking wildlife in uncontrolled environments

Wildlife monitoring has become increasingly important for conservation purposes, particularly in remote or hard-to-access areas where traditional methods, such as manual counting or camera-traps, are not practical. CV techniques have emerged as a powerful tool for automating this process by identifying and tracking animals through video footage. However, these approaches still face significant challenges due to the variability of lighting conditions, pose changes and backgrounds present in realistic scenarios. To overcome these limitations, the primary objective was to develop new techniques to identify animals in images and videos captured in natural habitats with minimal supervision. To achieve this, we propose a two-step approach, in which we proceed by initially introducing large-scale self-supervised learning solutions that use “label-free” images to learn generic characteristics applicable to different scenarios, followed by a possible finalizations applied using a limited number of labeled datasets adapted to particular species or environments. Our overall goal is to reduce the reliance on expensive manual labeling while enabling efficient deployment of state-of-the-art templates for real-world use cases. We expect the proposed solutions to provide better performance than current approaches that rely solely on fully supervised training or only unimodal feature extraction. Additionally, we plan to evaluate robustness against common sources of uncertainty faced by field operators or autonomous systems that collect media assets, such as variable lighting, occlusion, motion blur, etc. Finally, by sharing the knowledge gained in the course of this endeavor with a wider audience spanning multiple disciplines (particularly in the world of biology), we hope to stimulate a thoughtful dialogue about the potential ethical implications of the widespread adoption of intelligent monitoring equipment in unregulated environments.

Advisor: Niki Martinel

2) Machine Learning for decision support in the interpretation of images in anatomy

Microscope images can be acquired with special scanners, which produce images – called digital slides or WSI – at typical resolutions of 0.2-0.5 micron/pixel, on samples in the order of several squared mm-cm. The result is Gpixel images, rich in information which, precisely because of the size of the images, is still not fully exploited today. For the same reason, their systematic digitization is still rarely carried out, even if some laboratories or entire regional networks of laboratories are starting full digitization processes. It is a sector whose strong development began late compared to other medical specialties precisely due to the size of the images to be treated, which made their processing too complex for a long time, but which is now starting to have results of scientific interest both from the point of informatics and clinical point of view.

Advisor: Vincenzo Della Mea

Additional information

See the official PhD page at the University of Udine web site for calls and admission.

A list of the research topics of the PhD members is detailed here.