Towards Facilitative LLM
Staff - Faculty of Informatics
Date: 14 October 2024 / 13:30 - 14:30
USI East Campus, Room D1.15
Speaker: Prof. Hideo Joho, Institute of Library, Information and Media Science University of Tsukuba, Japan
Abstract: Most research on Large Language Models (LLMs) emphasizes automating human tasks, simulating users, or facilitating human-AI interaction, with less attention given to improving human-human communication. In this talk, I introduce the concept of a "Facilitative LLM," where an AI agent bridges knowledge gaps between individuals. I will begin by reviewing research on collaborative search and the key insights it provides. Then, I will demonstrate how information needs can be modeled and extracted from information seeking conversations. Finally, I will explore the broader implications of this approach and outline future directions for Facilitative LLMs, along with recent advances in conversational search systems and their evaluation frameworks.
Biography: Hideo Joho is a Full Professor at the Institute of Library, Information, and Media Science, University of Tsukuba, and an Honorary Research Fellow at the School of Computing Science, University of Glasgow (until March 2025). His research interests include human information interaction, interactive information retrieval, lifelogging, conversational search, and collaborative search. He has served as Program Co-Chair for iConference 2024 and CHIIR 2021/2019, General Co-Chair for SIGIR 2017, Associate Editor of Information Processing and Management (2014–16), and is a co-founder and the first chair of the Tokyo ACM SIGIR Chapter.
Host: Prof. Fabio Crestani