Takayuki Osogami

Short Bio

Takayuki Osogami is a senior technical staff member and the manager of the mathematical sciences group at IBM Research - Tokyo.  He is currently leading global research projects on reinforcement learning, industrial applications of reinforcement learning, automation of decision optimization with reinforcement learning,  multi-agent reinforcement learning, and integration of learning and game theory.  He was a group leader of a governmental project supported by Core Research for Evolutionary Science and Technology, Japan Science and Technology Agency during 2013-2019, where his group developed and applied  theory of sequential decision making, human behavior modeling, and neuromorphic computing.  He received his Ph.D. in Computer Science from Carnegie Mellon University in August 2005, and a B.Eng. degree in Electronic Engineering from the University of Tokyo in 1998.  He was selected as one of the Researchers with Nice Step 2020 by the National Institute of Science and Technology Policy (NISTEP) of the Ministry of Education, Culture, Sports, Science and Technology.

1998年よりIBM東京基礎研究所に勤務。2019年より同所シニア・テクニカル・スタッフ・メンバー、2020年より同所数理科学グループ担当。現在は、強化学習、強化学習の産業応用、強化学習による意思決定の自動化、マルチエージェント強化学習、強化学習とゲーム理論の融合にかかわるグローバル・プロジェクトをリード。2013年から2019年まで、JST-CRESTの研究課題「ディープナレッジを価値につなげるための意思決定最適化技術」を主たる共同研究者としてリード。1998年東京大学工学部電子工学科卒業。2005年カーネギーメロン大学コンピュータサイエンス学部コンピュータサイエンス学科博士号取得。科学技術への顕著な貢献2020（ナイスステップな研究者）。2010年日本オペレーションズ・リサーチ学会文献賞奨励賞受賞。2015年待ち行列研究部会論文賞受賞。

Research Overview

My research seeks to answer the question of how to make good decisions under uncertainties, where humans tend to make irrational decisions.  Uncertainties come up everywhere particularly when we make decisions in consideration of the future or other decision makers.  Human intuitions often fail under such uncertainties, and good decisions must be made on the basis of stochastic modeling, analysis, and optimization in a formal manner by fully exploiting available data.

My recent focus has been on sequential decision making under uncertainties, where one chooses actions iteratively in consideration of the benefit in the long term.  Time-series learning has also been a focus as a building block or a subproblem of sequential decision making.  A representative work in the context of time-series learning is on the dynamic Boltzmann machine (DyBM).  Sequential decision making under uncertainties becomes particularly difficult when we take into account risk or multiple agents, and much of my research deals with those difficulties.

As a researcher at IBM, I have a strong desire to develop new technologies, from my research, that will have major impacts on business, society, or science.  I believe that the role of a researcher is to solve problems that need to be solved but cannot be solved with existing technologies and to do so by developing new technologies.

Publications & Talks & Patents

Articles on My Work / Myself


Academic / Professional Services

Member of Governmental or Academic Committee

Member of Steering Committee

Member of Senior Program Committee or Area Chair

Member of Program Committee or Reviewers

Member of editorial board

Past activities

Honors / Grants