PhD Programme in Computer Science and Mathematics

• three places with University scholarship
• seven places Notice D.M. 630/24

1.     Notice: D.M.630/24   Project title and scholarship theme: Semantically enriched neural-symbolic methods for Knowledge Graph Refinement and Explanation

-         Research Activity Description: The project aims to develop neuro-symbolic solutions, capable of integrating data-driven methods and model-driven solutions (such as knowledge representation, knowledge graphs, complex reasoning and inference), as well as providing explanation for the solved tasks. The project will be focused on the development of neuro-symbolic methods with the aim of building robust Machine Learning models with the ability to represent, reason and explain numerical methods, characterized by high scalability on large amounts of data. The research will be particularly focused on the development of neuro-symbolic methods for: a) Knowledge Graph (KG) refinement tasks; d) providing explanations for the provided solutions. Improving the quality of KGs constitutes a fundamental element for improving the results of the numerous tasks currently solved with the massive usage of KGs. The research that will be developed will be functional to multiple disciplines, considering the use of KG in various fields such as e-commerce, web search, legal, medical. In particular, the possibility of verticalisation in the e-commerce and tourism sectors will be considered.

-         Co-financing company/body: Sidea Group S.r.l.

-         Period abroad: 6 months

-         Period at the Institution/Company: 6 months

 

2.     Notice: D.M.630/24

-         Project title and scholarship theme: Development and application of Generative Artificial Intelligence models based on the Symbiotic AI paradigm  to support the Public Administration 

-         Research Activity Description: The project aims to develop tools to support Public Administration using data-driven approaches and Artificial Intelligence (AI) based on the Symbiotic AI paradigm. The goal is to create AI applications that serve as a bridge between Public Administrations and citizens, facilitating citizens' access to existing services and information and improving the efficiency of public employees' daily activities. A central aspect of the project is using generative AI techniques in public administration services, focusing on Public Procurement. This includes implementing technological processes to facilitate the acquisition of goods and services and monitoring procurement procedures through intelligent management of information collected from regional and national procurement documents.

-         Co-financing company/body: InnovaPuglia SpA

-         Period abroad: 6 months

-         Period at the Institution/Company: 6 months

 

3.     Notice: D.M.630/24

-         Project title and scholarship theme: Generative approaches based on artificial intelligence for the automatic extraction of knowledge from books and manuals to support practical teaching activities. 

-         Research Activity Description: The project aims to study generative Artificial Intelligence (AI) techniques to support didactic and laboratory training for students from elementary school to university. Currently, in schools and universities, virtual reality (VR) systems are increasingly used for training, allowing the acquisition of skills in a simulated environment. However, for these systems it is expensive and time-consuming to create content. The project intends to provide a work environment that identifies entities, relationships and actions, offering intelligent assistance during teaching activities in a VR environment, where contents are created with generative AI techniques. The goal is to propose symbiotic environments in which AI, VR and students actively collaborate. New generative AI methods will be proposed to create content from educational texts and to interact with students, suggesting corrections during laboratory-practical activities.

-         Co-financing company/body: MTM Project Srl

-         Period abroad: 6 months

-         Period at the Institution/Company: 6 months

 

4.     Notice: D.M.630/24

-         Project title and scholarship theme: Development and application of Artificial Intelligence methods for human capital management and enhancement

-         Research Activity Description: The research project exploits the capabilities of Large Language Models (LLMs) in HR to create individual productivity tools and new ways of analyzing HR data, supporting teams’ strategic planning. Focusing on personalized content generation, analysis of employee feedback and interpretation of HR data using Artificial Intelligence, the research will help companies in digital transformation and development of digital skills needed for the modern labor market, integrating advanced technical and management skills.

-         Co-financing company/body: Talentia Software 

-         Period abroad: 6 months

-         Period at the Institution/Company: 6 months

 

5.     Notice: D.M.630/24

-         Project title and scholarship theme: Systems and techniques of Retrieval Augmented Generation for the exploitation of information resources in the industrial domain.

-         Research Activity Description: Many machine learning techniques (e.g., embedding, deep-memory, vector store, etc.) are being actively studied in academia and industry for the improvement of retrieval augmented generation systems. In this context, the specific objectives of the research will be: the development and application of benchmarks to characterize and quantify the effectiveness of existing systems based on the application context; and the development of innovative approaches or improvement of RAG systems based on specific quality, cost, and time objectives posed by the different industrial reference contexts (digital health, logistics, document management, etc.).

-         Co-financing company/body: Asclepyus srl

-         Period abroad: 6 months

-         Period at the Institution/Company: 6 months

 

6.     Notice: D.M.630/24

-         Project title and scholarship theme: CSS - Cyber Social Security: personal social security as an individual and as part of a community

-         Research Activity Description: The aim of the project is the analysis, design and development of solutions to support the Cyber Social Security, i.e. the social security of the person as an individual and as part of a community, adopting tools defined as social sensors, AI and adopting measures aimed at the safety of the instruments themselves. This objective can be achieved through the identification, response and prevention of certain events such as cases of violence, crimes and hate speech. The security of the tools used consists in identifying and protecting them from attacks aimed at compromising their functionality and reliability, known in the literature as "Data Poisoning" attacks.

-         Co-financing company/body: Cybersecurity srl

-         Period abroad: 6 months

-         Period at the Institution/Company: 6 months

 

7.     Notice: D.M.630/24

-         Project title and scholarship theme: Conversational agents and large language models to support services, products, and processes based on the Symbiotic AI paradigm.

-         Research Activity Description: The research project aims to explore the use of conversational agents and large language models, including multimodal ones, to support innovative services, products, and processes based on the Symbiotic AI paradigm, with the goal of improving user experience and operational efficiency across various sectors. The focus will be on applying Symbiotic AI technologies in contexts such as e-commerce, customer support, online education, and more. The core of the project will be the integration of advanced conversational agents and large multimodal models to optimize user-service interactions, facilitating information access and task execution in an intuitive and efficient manner.

-         Co-financing company/body: QuestIT S.r.l

-         Period abroad: 6 months

-         Period at the Institution/Company: 6 months

 

• six scholarships co-financed with companies

1.     n. 1 Scholarship co-funded by company: Asclepyus srl.

-         Project title and scholarship theme: Integration of analytical, heuristic and AI Generative models for primary prevention and treatment of disabling diseases.

-         Research Activity Description: The natural history of many disabling or degenerative diseases can be effectively described by mathematical (e.g., differential equations) and heuristic (e.g., statistical methods and machine learning) methods. The research will be concerned with combining the two approaches, including the use of generative AI techniques for synthesizing and comparing time series, quantitative and qualitative information about patients, with specific objectives such as analysis of data and time series; derivation of families of analytical models oriented toward personalized and precision medicine; development of effective diagnostic and predictive ML models complemented by analytical models; advanced use of retrieval augmented generation and generative AI techniques.

-         Period abroad: 6 months

-         Period at the Institution/Company: 6 months

 

2.     n. 1 Scholarship co-funded by company: InnovaPuglia SpA

-         Project title and scholarship theme: Predictive AI in Value Based Procurement (VBP).   

-         Research Activity Description: The aim of the project is the application of predictive Artificial Intelligence techniques in the field of public administration services, particularly in Public Procurement, in order to improve administrative processes with advanced Machine Learning tools that reference the information assets available through Empulia and available databases to build innovative pathways for research, design, development and validation of data-driven and user-centered solutions.  With a research effort thus geared toward Symbiotic AI specific to Process Mining, a powerful tool can be developed for the development of a value-based approach to the procurement of innovative technologies and services, particularly in healthcare.  

-         Period abroad: 6 months

-         Period at the Institution/Company: 6 months

 

3.     n. 1 Scholarship co-funded by company: Asclepyus S.r.l.

-       Project title and scholarship theme: Development of innovative anomaly detection techniques and application to patient health monitoring and early disease sign detection

-       Research Activity Description: The research program aims to develop advanced anomaly detection techniques also involving machine learning and artificial intelligence techniques, with applications in the field of Digital Health. The ultimate goal is to provide innovative mathematical tools useful for formulating predictive and diagnostic algorithms capable of improving disease management and personalized healthcare. In more detail, the project objectives include: researching numerical techniques to identify anomalous patterns in clinical and biometric datasets, enabling the timely detection of early signs of diseases; integrating outlier detection techniques to improve the predictive power of linear and non-linear models, possibly including machine learning and neural network-based techniques, in the presence of noisy data; Studying mathematical tools to support decision-making processes for more accurate diagnoses, real-time monitoring of patients’ condition evolution, and planning of personalized preventive interventions.

-       Period abroad: 6 months

-       Period at the Institution/Company: 6 months

 

4.     n. 1 Scholarship co-funded by company: AI Venture Builder

-        Project title and scholarship theme: Advanced generative models for conversion between images, text and audio

-       Research Activity Description: The research project aims to develop an advanced generative model that transforms visual, textual and audio signals, overcoming the limitations of current systems. Using advanced deep learning techniques and multimodal data analysis, the system will be able to interpret and translate visual features into detailed audio signals, transform audio signals into accurate textual descriptions, and generate realistic images from audio signals. Validation will be done by testing the quality and relevance of the outputs. The project aims to improve interaction and content representation, with therapeutic applications and increased accessibility for people with visual impairments.

-       Period abroad: 6 months

-       Period at the Institution/Company: 6 months

 

5.     n. 1 Scholarship co-funded by company: Digital Innovation S.r.l.

-       Project title and scholarship theme: AI methodologies for sustainability assessment of public administration processes

-       Research Activity Description: The research project aims to analyze, design and develop artificial intelligence (AI)-based solutions for process optimization in public administration (PA). Through the analysis of complex data and the generation of predictive models, the project intends to solve specific problems in local government, leveraging generative AI to improve efficiency, transparency and accessibility of services with a focus on the sustainability of the entire process. The system will provide benefits to citizens in terms of streamlined services, quick response times, transparent communication and improved accessibility, and for officials in terms of decision support, task automation, error reduction and increased efficiency.

-       Period abroad: 6 months

-       Period at the Institution/Company: 6 months

 

6.     n. 1 Scholarship co-funded by company:  AI² – Applicazioni di Ingegneria ed Informatica srl

-       Project title and scholarship theme: SymbioticRAIL2 – Symbiotic AI techniques supporting design and maintenance of railway systems

-       Research Activity Description: Rail systems are increasingly characterized by the wide variety of data involved in the different phases of the life cycle: formalization of requirements to design, verification of data consistency and expected operation, and diagnostics (for rapid fault prevention and analysis or in support for predictive maintenance). The risks of loss of quality and increased costs brought about by the growth in the complexity and quantity of information managed lead to the need for specific professional figures who are aware of the form and semantics of data. The activity is aimed at designing tools to support railway engineering through Data Driven approaches based on the Symbiotic AI paradigm.  It is focused on enhancing the capabilities of domain experts and designers to produce and (re)use knowledge in the different phases of the systems life cycle, thus nurturing a symbiotic process of training/learning and explanation underlying the use of AI tools.

-       Period abroad: 6 months

-       Period at the Institution/Company: 6 months