Fabrizio Silvestri is a Professor at the Department of Computer, Automatic and Management Engineering (DIAG) of Sapienza University of Rome, where he leads the RSTLess research group focused on Robust, Safe, and Transparent Deep Learning. His research spans Web Information Retrieval, Natural Language Processing, Machine Learning, and Deep Learning with emphasis on Retrieval Augmented Generation systems. Current projects investigate document chunking strategies in RAG to minimize distraction from irrelevant information, "backtracing" user query motivations, and fundamental parallels between Chain of Thought reasoning and information retrieval processes in large language models. Professor Silvestri actively supervises bachelor and master thesis students through structured projects in NLP and RAG, including investigations into chunking methodologies, query motivation analysis, and internal reasoning mechanisms. The RSTLess group maintains an active research agenda centered on developing transparent and robust deep learning frameworks while addressing practical challenges in information access systems.
