Hadas Raviv

Hadas is a Research Scholar at the Princeton Neuroscience Institute, where she works in the Hasson Lab. Her research focuses on how language is learned from naturalistic human experience and how it is represented in the human mind. She is especially interested in how language allows us to represent the world through perspectives shaped by our individual experiences and evolving over time, while also giving us a powerful means to shape, reinterpret, and communicate those experiences.

Despite their remarkable language capabilities, large language models and modern AI systems operate at scales and under training conditions that differ sharply from a child’s natural experience. This contrast motivates Hadas’s use of AI to investigate how language can be learned under more human-like conditions. Her work combines cognitively motivated training protocols with ultra-dense, real-world data that capture natural learning environments, along with evaluation methods designed to reflect human learning over realistic developmental timescales. She is currently focused on early lexical and syntactic learning, particularly on how children acquire language through everyday speech.

Hadas earned her PhD in Information Retrieval from the Technion—Israel Institute of Technology. She holds a BSc in Physics and Mathematics from the Hebrew University of Jerusalem and an MSc in Physics from Tel Aviv University. Her work across industry and academia has consistently centered on questions about language, learning, and representation.