Keynote Speaker I
Prof. Tetsuya Sakai, Waseda University, Japan
Tetsuya Sakai is a professor (2015–) in the Department of Computer Science and Engineering at Waseda University, Japan, and serves as Dean of the Center for Data Science (2026–). He is also a General Research Advisor at Naver Corporation, Korea (2021–), and a visiting professor at the National Institute of Informatics, Japan (2015–). He joined Toshiba in 1993 and received his Ph.D. from Waseda University in 2000 while continuing his work there. From 2000 to 2001, he conducted research under the supervision of the late Karen Sparck Jones at the Computer Laboratory, University of Cambridge. He later served as Director of the Natural Language Processing Lab at NewsWatch, Inc. (2007), and joined Microsoft Research Asia in 2009 before moving to the Waseda faculty in 2013. At Waseda, he served as Associate Dean (IT Strategies Division) from 2015 to 2017 and as Department Head from 2017 to 2019. He is an ACM Distinguished Member and an IPSJ Fellow. He was inducted into the SIGIR Academy in 2023 and received the Keith van Rijsbergen Award in 2026. Since 2021, he has served as Senior Associate Editor and Acting Editor-in-Chief of ACM Transactions on Information Systems.
Speech: Evaluating the Confidence of Retrieval-Augmented Generation Systems
Conversational search systems—particularly those built on Retrieval-Augmented Generation (RAG)—should express high confidence when their answers are correct, enabling users and downstream applications to rely on them. Conversely, they should signal uncertainty when they are not confident; in many cases, responding with “I don’t know” is preferable to providing an incorrect answer. This keynote presents the latest findings from a new evaluation task to be completed at NTCIR‑19 (the 19th NII Testbeds and Community for Information access Research conference), called R2C2: RAG Responses—Confident and Correct?. R2C2 is designed to assess the ability of RAG systems to produce appropriate confidence scores alongside their answers. Twelve research teams from eight countries and regions have registered to participate in the task.
Keynote Speaker II
Assoc. Prof. Pengfei Zhang, Chengdu University of Traditional Chinese Medicine (CDUTCM), China
Dr. Pengfei Zhang (张鹏飞) is currently an Associate Professor with the School of Intelligent Medicine, Chengdu
University of Traditional Chinese Medicine (CDUTCM). He received his Master Degree (Science) from the School
of Mathematical Sciences, Guangxi Minzu University, China, in June 2019. He obtained his Ph.D. degree from
the School of Computing and Artificial Intelligence, Southwest Jiaotong University (SWJTU), Chengdu, in June
2023.
He is the team leader of the Multi-modal Medical Intelligence & Multi-granularity Cognitive Computing Laboratory
(MAGIC Lab), CDUTCM. His publication topics include complex medical data processing (e.g., knowledge graph,
feature selection and anomaly detection) and multi-source information fusion. His broader research interests
include medical vertical large language model, medical data analysis, multi-modal learning, and the application
of artificial intelligence in Traditional Chinese Medicine (TCM), with particular emphasis on brain science and
acupuncture. He has authored over 60 peer-reviewed publications in high-impact journals and conferences, while
leading 4 competitive research projects as Principal Investigator (PI) funded by national and provincial grants.