|
Massimiliano Altieri (XXXVII Cycle)
|
Project Code: CN00000041
|
CUP: H93C22000430007
|
Adaptive deep learning methods for multiple time series analysis
Recently, additional attention has been given to collection and analysis of data in the form of time series. In fact, time series can be generated in multiple domains (e.g. sensor data analysis, system logs, cybersecurity) and require specific methods that are able to face all the challenges coming from the specific data organization. This project aims to study and propose novel adaptive and robust machine learning methods for time series analysis, with the purpose of tackling several fundamental challenges, that still represent an obstacle and require further investigation, such as extraction of spatio-temporal autocorrelation, scalability and explainability.
|
Project: Piano Nazionale di Ripresa e Resilienza, Missione 4 “Istruzione e ricerca” Componente 2 Investimento 1.4 “Potenziamento di strutture di ricerca e creazione di campioni nazionali di R&S su alcune Key Enabling Technologies” finanziato dall’Unione europea – NextGenerationEU, per il Progetto “National Center for Gene Therapy and Drugs based on RNA Technology” (codice progetto CN00000041_CUP H93C22000430007)
|
Natalino Borgia (XXXVIII Cycle)
|
Project Code: CN00000013
|
CUP: H93C22000450007
|
Nonlinear PDEs prospective on a Changing Environment
The natural processes that are the triggers for natural disasters are classified into several categories: geophysical, hydrological, meteorological, climatological, biological, extraterrestrial. Mathematical sciences can help in describing, predicting, monitoring and responding to such events and mitigating their effects. This research project will focus on some Nonlinear Partial Differential Equations which are related to nonlinear phenomena and having some impact on the Environment. A first goal is the study of quasilinear elliptic equations and systems, involving p-laplace operator or p-area operator.
|
Project: Piano Nazionale di Ripresa e Resilienza (PNRR) Missione 4, Componente 2 “Dalla Ricerca all'impresa” – Investimento 1.4 “Potenziamento strutture di ricerca e creazione di "Campioni Nazionali di R&S” su alcune Key Enabling Technologies” finanziato dall’Unione Europea – NextGenerationEU (DM MUR n.3138 del 16.12.2021) – Progetto Centro Nazionale HPC, Big Data e Quantum Computing, SPOKE 5 “Environment & Natural Disasters”
|
Veronica Buttaro (XXXIX Cycle)
|
Project Code: PE00000013
|
CUP: H97G22000210007
|
Machine learning model recycling for biomedical applications
The rapid development of AI and ML has revolutionized many fields, including medicine and biology, but it has also led to increased energy consumption and CO2 emissions, especially during the training phases of AI models. My research project focuses on enhancing the sustainability of AI by developing strategies to reduce computational resources and improve data management efficiency. Key approaches include smart reuse of AI models, adapting pretrained models for new contexts, and using ensemble techniques to combine models for better predictions. The project will apply these sustainable AI methods in the biomedical domain, aiming to improve early disease diagnosis, identify new biomarkers, and explore drug discovery and repositioning through transfer learning. This work seeks to reduce the environmental impact of AI while making it more efficient and affordable.
|
Project: Piano Nazionale di Ripresa e Resilienza (PNRR) – Missione 4 “Istruzione e ricerca” – Componente 2 “Dalla ricerca all’impresa” Investimento 1.3, finanziato dall’Unione europea – NextGenerationEU. – Partenariato Esteso “Future ArtificialIntelligence Research” FAIR (Spoke 6: Symbiotic AI)
|
Miriana Calvano (XXXIX Cycle)
|
Project Code: PE00000013
|
CUP: H97G22000210007
|
Study and application of methods and guidelines for human-centered design of symbiotic systems
Artificial Intelligence (AI) systems rapidly propagate across various domains, revolutionising industries, and augmenting human capabilities. There is a need to design high-quality AI systems that focus on the users' priorities and avoid potential unethical behaviours. In the current scenario, Human-Computer Interaction (HCI) and AI are not separate fields, but they contaminate each other. To create high-quality Symbiotic AI (SAI) systems and to foster the human-AI symbiosis the human-centered design approach should be adopted.
This project aims at defining metrics to evaluate the symbiotic relationship and the quality of SAI systems considering both the user's and AI system's performance. This approach can be useful to better understand how SAI systems can be designed and implemented to most effectively serve the humans’ needs.
|
Project: European Union - Next Generation EU: NRRP Initiative, Mission 4, Component 2, Investment 1.3 -- Partnerships extended to universities, research centers, companies, and research D.D. MUR n. 341 del 15.03.2022 -- Next Generation EU (PE0000013 -- ``Future Artificial Intelligence Research -- FAIR'' - CUP: H97G22000210007)
|
Faiza Hasin (XXXVIII Cycle)
|
Project Code: CN00000041
|
CUP: H93C22000450007
|
Design of explainable methods for predicting on-target and off-target effects of gRNA sequences in CRISPR/Cas9 Base Editing.
The CRISPR/Cas9 system is a versatile genome-editing tool used for gene knockout and base editing. However, off-target effects pose challenges due to DNA repetition and individual genetic variation, leading to unpredictable outcomes. Predicting off-target effects is crucial for reducing risks and obtaining accurate results. This project aims to develop explainable methods to predict on-target and off-target effects in CRISPR/Cas9 Base Editing, focusing on optimizing gRNA sequence design for CRISPR screens on human cell lines to study gene editing's impact on cancer treatment and genetic diseases. The project includes addressing uncertainty and providing models with explanatory capabilities and uncertainty management. The expected outcomes include a methodology, predictive models, open-source software, and a protocol for biotechnologists to optimize gRNA sequences. The findings will be disseminated through publications in conference proceedings and archival journals.
|
Project: Piano Nazionale di Ripresa e Resilienza (PNRR) Missione 4, Componente 2 “Dalla Ricerca all'impresa” – Investimento 1.4 “Potenziamento strutture di ricerca e creazione di "Campioni Nazionali di R&S” su alcune Key Enabling Technologies” finanziato dall’Unione Europea – NextGenerationEU (DM MUR n.3138 del 16.12.2021) – National Center for Gene Therapy and Drugs based on RNA Technology, SPOKE 7 "BioComputing"
|
Alessandro Petruzzelli (XXXIX Cycle)
|
Project Code: PE00000013
|
CUP: H97G22000210007
|
Understanding human needs in Symbiotic AI: towards Symbiotic Recommender Systems
In my Ph.D. research, I will explore Symbiotic Conversational Recommender Systems (SCRSs) powered by Large Language Models (LLMs) to align the recommendation process with symbiotic AI principles. CRSs leverage user interaction through natural language processing to provide personalized recommendations collaboratively and interactively during the recommendation process. While LLMs show promise for traditional RS tasks, integrating them into an open-ended CRS is challenging. My Ph.D. proposes a novel approach to adapt LLMs for CRS by focusing on three key areas: Adapting LLMs for Recommendation Tasks; integrating the personal information of the user to adapt the response strategy; and guaranteeing symbiotic interaction thought a multimodal model. Initial experiments highlight the limitations of LLM performance in areas not well-represented in their training data.
|
Project: Piano Nazionale di Ripresa e Resilienza (PNRR) – Missione 4 “Istruzione e ricerca” – Componente 2 “Dalla ricerca all’impresa” Investimento 1.3, finanziato dall’Unione europea – NextGenerationEU. – Partenariato Esteso “Future Artificial Intelligence Research” FAIR (Spoke 6: Symbiotic AI)
|
Sileshi Nibret Zeleke (XXXVIII Cycle)
|
Project Code: PE0000015
|
CUP: H33C22000680006
|
Explainable computational methods for the analysis of signals (time series and images) related to biometric parameters.
In recent years, the explainability of AI models and the privacy preservation of health data using federated learning have gained significant attention in the digital health domain. My PhD project aims to utilize explainable AI and federated learning techniques in biometric signal processing to provide a privacy-preserving, interpretable approach. We focus on making this approach usable for conditions related to the aging population, such as cardiovascular disease detection, neurological disorder monitoring, mental health, and in-home monitoring of the elderly. Additionally, we concentrate on continuous health state monitoring and interventions in various aspects of older people's lives while preserving their data privacy. Biometric parameters collected using sensors in-body, on-body, near-body, and far-body serve as data sources to monitor and allow continuous observation. Our general objective is to design an interpretable federated learning-based system for in-home monitoring of elderly healthcare.
|
Project: Piano Nazionale di Ripresa e Resilienza (PNRR) Missione 4, Componente 2 “Dalla Ricerca all'impresa” – Investimento 1.3, Partenariati estesi alle università, ai centri di ricerca, alle aziende per il finanziamento di progetti di ricerca di base, DD MUR 341 del 15/03/2022. Progetto: Age-It: Ageing well in an ageing society - A novel public-private alliance to generate socioeconomic, biomedical and technological solutions for an inclusive Italian ageing society.
|
|