言語の枠を超えて：大規模言語モデルの可能性と挑戦

ID:C-21 Concurrent Sessions

Beyond the boundaries of language: The Potential and Challenges of Large Language Models 言語の枠を超えて：大規模言語モデルの可能性と挑戦

開催日時：2025年10月2日（木曜日） / 13:30 - 16:30 Onsite Date：October 2, 2025 / 13:30 - 16:30 Onsite 開催場所：Hall200 Room：Hall200

Tsukuba Institute for Advanced Research, University of Tsukuba 筑波大学高等研究院
Center for Artificial Intelligence Research 筑波大学人工知能科学センター

Research and development in AI, especially large language models (LLMs), has progressed rapidly in recent years, and it has now become an essential core technology in various fields, far beyond the scope of natural language processing. In this session, using LLMs as a common thread, young researchers from various fields and AI researchers will discuss what kind of frontier research can be expected and realized in areas where LLMs intersect with each field. Following topic presentations by researchers from each field, there will be a discussion involving the audience as well.

AI、特に大規模言語モデル（LLM）の研究開発はここ数年で急激に進展し，今では自然言語処理の枠を大きく超えて，様々な分野における不可欠な中核技術となっています．本セッションではLLMを横串として，各分野の若手研究者とAI研究者が，LLMと各分野がクロスする領域で，どのようなフロンティア研究が期待・実現できるのか、各分野の研究者の話題提供に続き，会場も交えたディスカッションを行います。

Speakers

Kyosuke Nagata 永田恭介 President, University of Tsukuba
Director General, Tsukuba Institute for Advanced Research (TIAR) 筑波大学長、高等研究院（TIAR）研究院長
Kazuhiro Fukui 福井和広 Director, Center for Artificial Intelligence Research, Tsukuba Institute for Advanced Research (TIAR), University of Tsukuba 筑波大学高等研究院（TIAR）、人工知能科学センター長
Roberto Cipolla Professor, Information Engineering, University of Cambridge Managing Director, Cambridge Research Laboratory, Toshiba Europe Limited Professor, Information Engineering, University of Cambridge Managing Director, Cambridge Research Laboratory, Toshiba Europe Limited
Jerry Li Associate Professor, Paul G. Allen School of Computer Science & Engineering, University of Washington Associate Professor, Paul G. Allen School of Computer Science & Engineering, University of Washington
Hao Shen Postdoctoral Researcher, Baker Lab, Institute for Protein Design, University of Washington Postdoctoral Researcher, Baker Lab, Institute for Protein Design, University of Washington
Yasuhiko Igarashi 五十嵐康彦 Associate Professor, Institute of Systems and Information Engineering, Center for Artificial Intelligence Research, Hydrogen Boride Research Center, Tsukuba Institute for Advanced Research (TIAR), University of Tsukuba 筑波大学システム情報系准教授
高等研究院（TIAR）
人工知能科学センター
ホウ化水素研究センター
Yuji Hirono 広野雄士 Associate Professor, Institute of Systems and Information Engineering, University of Tsukuba 筑波大学システム情報系准教授
Pedro Valois 　 Doctoral Program in Computer Science, University of Tsukuba 筑波大学情報理工学位プログラム
Akiyoshi Fukamizu 深水昭吉 Director, Office for Strategy Promotion, Tsukuba Institute for Advanced Research (TIAR), University of Tsukuba 筑波大学高等研究院（TIAR）、戦略推進室長

Time Table

Time Title Speaker Affiliation
13:30 - 13:40

Opening Remarks 開会挨拶

Kyosuke Nagata 永田恭介

President, University of Tsukuba
Director General, Tsukuba Institute for Advanced Research (TIAR) 筑波大学長、高等研究院（TIAR）研究院長
13:40 - 13:50

Introduction to C-AIR and LLM Overview C-AIR及びLLM概要紹介

Kazuhiro Fukui 福井和広

Director, Center for Artificial Intelligence Research, Tsukuba Institute for Advanced Research (TIAR), University of Tsukuba 筑波大学高等研究院（TIAR）、人工知能科学センター長
13:50 - 14:20

Keynote Speech
Road to Physical Intelligence 基調講演
Road to Physical Intelligence

Roberto Cipolla Roberto Cipolla

Professor, Information Engineering, University of Cambridge Managing Director, Cambridge Resrach Laboratory, Toshiba Europe Limited Professor, Information Engineering, University of Cambridge Managing Director, Cambridge Resrach Laboratory, Toshiba Europe Limited
14:20 - 14:35

Invited Talk 1
The Delta Learning Hypothesis: Preference Tuning on Weak Data can Yield Strong Gains 招待講演１
The Delta Learning Hypothesis: Preference Tuning on Weak Data can Yield Strong Gains

Jerry Li Jerry Li

Associate Professor, Paul G. Allen School of Computer Science & Engineering, University of Washington Associate Professor, Paul G. Allen School of Computer Science & Engineering, University of Washington
14:35 - 14:50

Invited Talk 2
De novo design of programmed self-assembling protein filaments for function engineering 招待講演２
De novo design of programmed self-assembling protein filaments for function engineering

Hao Shen Hao Shen

Postdoctoral Researcher, Baker Lab, Institute for Protein Design, University of Washington Postdoctoral Researcher, Baker Lab, Institute for Protein Design, University of Washington
14:50 - 15:05

Invited Talk 3
LLMs as Knowledge Engines: Digitalizing Measurements and Designing Experiments under Data Scarcity 招待講演３
LLMs as Knowledge Engines: Digitalizing Measurements and Designing Experiments under Data Scarcity

Yasuhiko Igarashi 五十嵐康彦

Associate Professor, Institute of Systems and Information Engineering, Center for Artificial Intelligence Research, Hydrogen Boride Research Center, Tsukuba Institute for Advanced Research (TIAR), University of Tsukuba 筑波大学システム情報系准教授、高等研究院（TIAR）、人工知能科学センター、ホウ化水素研究センター
15:05 - 15:20

Invited Talk 4 招待講演４

Yuji Hirono 広野雄士

Associate Professor, Institute of Systems and Information Engineering, University of Tsukuba 筑波大学システム情報系准教授
15:20 - 15:35

Invited Talk 5
How to make Large Language Models? 招待講演５
How to make Large Language Models?

Pedro Valois Pedro Valois

Doctoral Program in Computer Science, University of Tsukuba 筑波大学情報理工学位プログラム
15:35 - 15:45

Break 休憩
15:45 - 16:25

Panel Discussion パネルディスカッション

All speakers 発表者全員
16:25 - 16:30

Closing Remarks 閉会挨拶

Akiyoshi Fukamizu 深水昭吉

Director, Office for Strategy Promotion, Tsukuba Institute for Advanced Research (TIAR), University of Tsukuba 筑波大学高等研究院（TIAR）、戦略推進室長