Research
My research centres on Natural Language Processing and related AI technologies, with a focus on language models, semantics, multimodality, and domain-specific applications.
I am especially interested in the development and evaluation of NLP systems for real-world scenarios, including misinformation detection, affective computing, information extraction, legal NLP, domain adaptation, and intelligent document analysis.
Research areas
- Trustworthy NLP and responsible AI
- Language resources and evaluation
- Semantics and Interpretability
- Domain adaptation, continual learning, and multimodality
- NLP and LLMs for Software Engineering
Selected projects with leading roles
CLMs4BPO — Conditional Language Models for Business Process Optimization
Scientific Coordinator.
Project funded by Regione Toscana for the activation of research fellowships.
The project focused on the automatic generation of text conditioned by explicit metadata, using generative AI techniques and instruction-tuned Large Language Models.
PRIN 2022 — Psychologically-tailored approaches to Debunk Fake News
Work Package Leader.
The project studies fake news recognition and personalized debunking strategies based on the psychological and behavioural characteristics of individuals.
The work combines misinformation detection, NLP, and user-aware approaches to debunking.
SEM — Sem Il Chattadino
Scientific Coordinator.
Project funded by Regione Toscana POR CreO FESR 2014–2020.
The project aimed at developing a chatbot for the Public Administration domain, bringing together CoLing Lab and partner companies.
Ongoing and recent research projects
EMERGE — Emergent Awareness from Minimal Collectives
European project funded under the EIC programme.
The project investigates emergent forms of collaborative awareness in collectives of minimal artificial agents, focusing on how simple agents can develop representations of each other, their environment, and cooperative behaviours.
My contribution concerns language-model updating and multimodal communication in virtual agents.
FAIR — Future Artificial Intelligence Research
The project addresses research questions, methodologies, models, technologies, and ethical and legal aspects for building Artificial Intelligence systems.
My contribution is connected to Spoke 1, coordinated by the University of Pisa, with a focus on Human-Centered Lifelong Learning for Complex Data.
TAILOR — Foundations of Trustworthy AI
European H2020 project on Trustworthy AI, integrating reasoning, learning, and optimization.
The consortium promotes a human-centred vision of responsible AI based on expertise, cooperation, and diversity.
PRIN 2022 PNRR — New Frontiers for a Participatory Criminal Justice Towards the Thresholds of Metaverse
The project investigates the impact of ICT technologies on judicial decision-making and on the conduct of remote legal proceedings.
Past projects at CoLing Lab
Giustizia Agile — PON 2014–2020
The project aimed to enhance the efficiency of judicial offices through technological innovation and organizational support.
The University of Pisa group worked on information extraction from legal documents and on the development of tools for document consultation.
Event Extraction for Fake News Detection
The project focused on the development of a system for fake news detection based on graph representations of news events and actors.
MUSE — Multimodal Semantic Extraction
Project funded by Regione Toscana POR FSE 2014–2020.
The project focused on multimodal semantic analysis of texts and images, combining Natural Language Processing and Computer Vision techniques.
UBIMOL — Ubiquitous Massive Open Learning
Project funded by Regione Toscana POR FESR 2014–2020.
The project aimed at developing an e-learning platform enriched with NLP technologies for personalized learning.
Memorie di Guerra
The project applied NLP techniques to war memories and historical documents, with the goal of supporting historical research and preserving historical memory.
SEMPLICE — Semantic Instruments for Public Administrators and Citizens
Two-year project funded by Regione Toscana.
The project focused on NLP-based tools for knowledge management, information extraction, and opinion mining for local public administrations.