Takuma Seno

Ph.D, Research Scientist @ Sony Research

EXPERIENCES

  • Sony Research (2020-10 - Present)

    Research Scientist, working for Gran Turismo Sophy

  • Okinawa Institute of Science and Technology (2020-03 - 2020-07)

    Research Assistant, worked for a custom backpropagation library: Proppo

  • Sony Corporation (2018-11 - 2020-07)

    Part-time Resarch Engineer, worked for Super-human Gran Turismo AI and Neural Network Libraries

  • Ghelia (2018-04 - 2020-08)

    Part-time Research Engineer, worked for reinforcement learning projects (e.g. robotics, numerical optimization)

  • Wantedly (2016-03 - 2018-11)

    Part-time Web Developer, worked for wantedly.com and recommendation algorithms

  • MatchinGood (2014-10 - 2016-04)

    Part-time Web Developer, worked for building Web systems

SELECTED PUBLICATIONS

  • Hojoon Lee*, Dongyoon Hwang*, Donghu Kim, Hyunseung Kim, Jun Jet Tai, Kaushik Subramanian, Peter R. Wurman, Jaegul Choo, Peter Stone, Takuma Seno, "SimBa: Simplicity Bias for Scaling Up Parameters in Deep Reinforcement Learning", ArXiv Preprint, 2024

    PaperProject PageCode

  • Miguel Vasco*, Takuma Seno*, Kenta Kawamoto, Kaushik Subramanian, Peter R. Wurman, Peter Stone, "A Super-human Vision-based Reinforcement Learning Agent for Autonomous Racing in Gran Turismo", 1st Reinforcement Learning Conference (RLC), 2024, Outstanding Paper Award on Applications of RL

    PaperProject Page

  • Paavo Parmas, Takuma Seno, Yuma Aoki, "Model-based Reinforcement Learning with Scalable Composite Policy Gradient Estimators", 40th International Conference on Machine Learning (ICML), 2023

    Paper

  • Takuma Seno, Michita Imai, "d3rlpy: An Offline Deep Reinforcement Learning Library", Journal of Machine Learning Research (JMLR), 315, 1-20, 2022

    PaperCode

  • Paavo Parmas, Takuma Seno, "Proppo: a Message Passing Framework for Customizable and Composable Learning Algorithms", 36th Conference on Neural Information Processing Systems (NeurIPS), 2022

    PaperCode

  • James MacGlashan, Evan Archer*, Alisa Devlic*, Takuma Seno*, Craig Sherstan*, Peter R Wurman, Peter Stone, "Value Function Decomposition for Iterative Design of Reinforcement Learning Agents", 36th Conference on Neural Information Processing Systems (NeurIPS), 2022

    PaperCode

  • Peter R Wurman, Samuel Barrett, Kenta Kawamoto, James MacGlashan, Kaushik Subramanian, Thomas J Walsh, Roberto Capobianco, Alisa Devlic, Franziska Eckert, Florian Fuchs, Leilani Gilpin, Piyush Khandelwal, Varun Kompella, HaoChih Lin, Patrick MacAlpine, Declan Oller, Takuma Seno, Craig Sherstan, Michael D Thomure, Houmehr Aghabozorgi, Leon Barrett, Rory Douglas, Dion Whitehead, Peter Dürr, Peter Stone, Michael Spranger, Hiroaki Kitano. "Outracing champion Gran Turismo drivers with deep reinforcement learning", Nature, 602, 223-228, 2022

    PaperProject Page

  • Ryuji Imamura, Takuma Seno, Kenta Kawamoto, Michael Spranger, "Expert Human-Level Driving in Gran Turismo Sport Using Deep Reinforcement Learning with Image-based Representation", NeurIPS Deep Reinforcement Learning Workshop, 2021

    Paper

SOFTWARE PROJCTS

EDUCATIONS

  • Keio University (2019-09 - 2023-09)

    Ph.D, Computer Science

  • Keio University (2018-04 - 2019-09)

    Master, Computer Science

  • Keio University (2014-04 - 2018-03)

    Bachelor, Computer Science

AWARDS

  • Outstanding Paper Award on Applications of RL received from Reinforcement Learning Conference (2024)

  • Mitou Super Creator certified by Information-technology Promotion Agency, Japan (IPA) (2021)

  • Mitou Creator selected by Information-technology Promotion Agency, Japan (IPA) (2020)

  • Best Presentation Award given at Annual Conference in Japanese Neural Network Society (2019)

  • Student Encouragement Award given at Annual Conference in Information Processing Society Japan (2018)

  • Student Encouragement Award given at Human-Agent Interaction Symposium (2017)