معرفی
Sandro Pezzelle is an Assistant Professor in Responsible AI at the Institute for Logic, Language and Computation (ILLC), Faculty of Science, University of Amsterdam. He focuses on language-mediated intelligence in humans and machines, combining insights from Natural Language Processing (NLP), Machine Learning, and Cognitive Science. His research examines computational semantics/pragmatics, language-vision models, cognitive mechanisms in human semantics, and LLM interpretability.
Education:
- PhD at CIMeC, University of Trento (supervised by Raffaella Bernardi)
- Postdoc in the DREAM ERC project led by Raquel Fernández
- 2018 Research Intern at SAP AI Research
Research Interests: Pezzelle explores how language models represent meaning, their visual reasoning capabilities, and their alignment with human cognitive processes. He investigates mechanisms enabling LLMs to adapt to interlocutors' styles and assesses their interpretability through behavioral and mechanistic lenses. His work bridges formal linguistics, cognitive science, and AI ethics.
Affiliations:
- Member of CERTAIN (Center for Explainable, Responsible AI)
- ELLIS Faculty member
- Board member of ACL SigSem
- Scientific Advisor at IVADO Labs
Recent Activities: Keynote speaker at MuMiCo 2024, CMCL Workshop (ACL 2024), and multiple international venues. Active in publishing groundbreaking work on LLM evaluation and interpretability, with recent preprints and peer-reviewed papers in ACL/EMNLP venues.
