• Search Research Projects
  • Search Researchers
  • How to Use
  1. Back to previous page

Human-like continual robot learning based on three-level computational energy cost regulation

Research Project

Project/Area Number 23K24926
Project/Area Number (Other) 22H03670 (2022-2023)
Research Category

Grant-in-Aid for Scientific Research (B)

Allocation TypeMulti-year Fund (2024)
Single-year Grants (2022-2023)
Section一般
Review Section Basic Section 61050:Intelligent robotics-related
Research InstitutionOsaka University

Principal Investigator

OZTOP Erhan  大阪大学, 先導的学際研究機構, 特任教授(常勤) (90542217)

Project Period (FY) 2022-04-01 – 2025-03-31
Project Status Completed (Fiscal Year 2024)
Budget Amount *help
¥14,950,000 (Direct Cost: ¥11,500,000、Indirect Cost: ¥3,450,000)
Fiscal Year 2024: ¥4,160,000 (Direct Cost: ¥3,200,000、Indirect Cost: ¥960,000)
Fiscal Year 2023: ¥4,550,000 (Direct Cost: ¥3,500,000、Indirect Cost: ¥1,050,000)
Fiscal Year 2022: ¥6,240,000 (Direct Cost: ¥4,800,000、Indirect Cost: ¥1,440,000)
KeywordsContinual Robot Learning / Multi-task Learning / Task switching / Learning Progress / Energy Regulation / Symbol Formation / Intrinsic Motivation / Cognitive Robotics / 計算エネルギーコスト / 継続的な行動学習 / 概念形成 / 内発的動機 / スキル移転 / Lifelong Robot Learning / Knowledge Transfer / Multitask Learning / Interleaved Learning
Outline of Research at the Start

生物にはエネルギーコストの制限があり,生涯を通じて効率的で有用な行動の学習と発達を可能にする.この制約は,ロボットにおいては,計算エネルギーコスト(CEC,computational energy cost)制約に対応すると考えられるが,ヒトと同様にこの制約がロボットの継続的な行動学習・発達にどのようにうまく機能するかが,本研究のテーマである.これを検証するために,ロボットの行動学習・発達機構として三層構造(エネルギーコスト最小化するニューラルネットワーク, CEC基づいた内発的動機, 概念形成)を想定する.各層でこの制約に基づく計算手法をロボットに実装され,ヒトに類似した学習行動が生成する.

Outline of Final Research Achievements

The project aims to develop mechanisms for multitask learning that resemble human learning, with a focus on computational efficiency. The scope was extended to address related challenges, including human-like reinforcement learning, active learning to reduce learning cost, and robot learning in social contexts where trust relates to energy savings. Key achievements include: (1) Generalizing skill transfer to be bidirectional, removing fixed source-target task roles; (2) Proposing an interleaved learning scheme enabling lifelong learning without fixed task order; (3) Developing a deep learning model with attention and energy-modulated learning progress (EMLP), achieving better multitask performance; (4) Designing a multitask RL framework with human-like interleaved learning and bidirectional skill transfer; (5) Introducing neuro-symbolic architectures to reduce planning cost; (6) Proposing prediction uncertainty as an intrinsic motivation signal to enhance sample efficiency.

Academic Significance and Societal Importance of the Research Achievements

本研究は、「学習進捗」と「神経的エネルギー」のバランスを取る人間類似のタスクスイッチング機構を導入することで、マルチタスク学習の性能向上と神経的コストの削減を両立させた点に学術的意義がある。
本プロジェクトの成果は、ロボットが自律的に「何を・いつ学習すべきか」を判断可能とすることにより、人間とロボットが共生する社会の実現に向けた基盤を築くものである。

Report

(3 results)
  • 2024 Final Research Report ( PDF )
  • 2023 Annual Research Report
  • 2022 Annual Research Report
  • Research Products

    (16 results)

All 2024 2023 2022 Other

All Int'l Joint Research (3 results) Journal Article (6 results) (of which Int'l Joint Research: 3 results,  Peer Reviewed: 3 results,  Open Access: 1 results) Presentation (6 results) (of which Int'l Joint Research: 6 results) Funded Workshop (1 results)

  • [Int'l Joint Research] Bogazici University/Ozyegin University(トルコ)

    • Related Report
      2023 Annual Research Report
  • [Int'l Joint Research] Tilburg University(オランダ)

    • Related Report
      2023 Annual Research Report
  • [Int'l Joint Research] Bogazici University/Ozyegin University(トルコ)

    • Related Report
      2022 Annual Research Report
  • [Journal Article] Diffusion Policies for Out-of-Distribution Generalization in Offline Reinforcement Learning2024

    • Author(s)
      Ada Suzan Ece、Oztop Erhan、Ugur Emre
    • Journal Title

      IEEE Robotics and Automation Letters

      Volume: 9 Issue: 4 Pages: 3116-3123

    • DOI

      10.1109/lra.2024.3363530

    • Related Report
      2023 Annual Research Report
    • Peer Reviewed / Int'l Joint Research
  • [Journal Article] Discovering Predictive Relational Object Symbols With Symbolic Attentive Layers2024

    • Author(s)
      Ahmetoglu Alper、Celik Batuhan、Oztop Erhan、Ugur Emre
    • Journal Title

      IEEE Robotics and Automation Letters

      Volume: 9 Issue: 2 Pages: 1977-1984

    • DOI

      10.1109/lra.2024.3350994

    • Related Report
      2023 Annual Research Report
    • Peer Reviewed / Int'l Joint Research
  • [Journal Article] Correspondence Learning Between Morphologically Different Robots via Task Demonstrations2024

    • Author(s)
      Aktas Hakan、Nagai Yukie、Asada Minoru、Oztop Erhan、Ugur Emre
    • Journal Title

      IEEE Robotics and Automation Letters

      Volume: 9 Issue: 5 Pages: 4463-4470

    • DOI

      10.1109/lra.2024.3382534

    • Related Report
      2023 Annual Research Report
    • Peer Reviewed / Open Access / Int'l Joint Research
  • [Journal Article] Trust in robot-robot scaffolding2023

    • Author(s)
      Kirtay Murat、Hafner Verena V.、Asada Minoru、Oztop Erhan
    • Journal Title

      IEEE Transactions on Cognitive and Developmental Systems

      Volume: 1 Issue: 4 Pages: 1-1

    • DOI

      10.1109/tcds.2023.3235974

    • Related Report
      2022 Annual Research Report
  • [Journal Article] DeepSym: Deep Symbol Generation and Rule Learning for Planning from Unsupervised Robot Interaction2022

    • Author(s)
      Ahmetoglu Alper、Seker M. Yunus、Piater Justus、Oztop Erhan、Ugur Emre
    • Journal Title

      Journal of Artificial Intelligence Research

      Volume: 75 Pages: 709-745

    • DOI

      10.1613/jair.1.13754

    • Related Report
      2022 Annual Research Report
  • [Journal Article] Multimodal reinforcement learning for partner specific adaptation in robot-multi-robot interaction2022

    • Author(s)
      Kirtay Murat、Hafner Verena V.、Asada Minoru、Kuhlen Anna K.、Oztop Erhan
    • Journal Title

      IEEE Proceedings on Humanoids 2022, Ginowan, Japan

      Volume: 1 Pages: 1-1

    • DOI

      10.1109/humanoids53995.2022.10000205

    • Related Report
      2022 Annual Research Report
  • [Presentation] Human-in-the-Loop Training Leads to Faster Skill Acquisition and Adaptation in Reinforcement Learning-based Robot Control2024

    • Author(s)
      Yilmaz D, Ugurlu B, Oztop E
    • Organizer
      18th IEEE International Conference on Advanced Motion Control (AMC2024), Kyoto, Japan
    • Related Report
      2023 Annual Research Report
    • Int'l Joint Research
  • [Presentation] A Model for Cognitively Valid Lifelong Learning2023

    • Author(s)
      Say H, Oztop E
    • Organizer
      IEEE International Conference on Robotics and Biomimetics (ROBIO 2023), Koh Samui, Thailand
    • Related Report
      2023 Annual Research Report
    • Int'l Joint Research
  • [Presentation] Developmental Scaffolding with Large Language Models2023

    • Author(s)
      Celik MB, Ahmetoglu A, Ugur E, Oztop E
    • Organizer
      23rd IEEE International Conference on Development and Learning (ICDL 2023), Macau, China
    • Related Report
      2023 Annual Research Report
    • Int'l Joint Research
  • [Presentation] Interplay between neural computational energy and multimodal processing in robot-robot interaction2023

    • Author(s)
      Kirtay M, Hafner VV, Asada M, Oztop E
    • Organizer
      23rd IEEE International Conference on Development and Learning (ICDL 2023), Macau, China
    • Related Report
      2023 Annual Research Report
    • Int'l Joint Research
  • [Presentation] Context Based Echo State Networks for Robot Movement Primitives2023

    • Author(s)
      Amirshirzad N, Asada M, Oztop E
    • Organizer
      32nd IEEE International Conference on Robot & Human Interactive Communication (RO-MAN) Busan, South Korea
    • Related Report
      2023 Annual Research Report
    • Int'l Joint Research
  • [Presentation] Bimanual rope manipulation skill synthesis through context dependent correction policy learning from human demonstration2023

    • Author(s)
      Akbulut Baturhan, Tuba Girgin, Mehrabi Arash, Ugur Emre, Oztop Erhan
    • Organizer
      IEEE International Conference on Robotics and Automation (ICRA2023)
    • Related Report
      2022 Annual Research Report
    • Int'l Joint Research
  • [Funded Workshop] IEEE IROS 2022 Workshop on “Lifelong Learning of High-level Cognitive and Reasoning Skills” (https://lifelongrobotics.github.io/)2022

    • Related Report
      2022 Annual Research Report

URL: 

Published: 2022-04-19   Modified: 2026-01-16  

Information User Guide FAQ News Terms of Use Attribution of KAKENHI

Powered by NII kakenhi