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

Base technology of "Social Expression" for dialogue robots

Publicly Offered Research

Project AreaStudies on intelligent systems for dialogue toward the human-machine symbiotic society
Project/Area Number 20H05576
Research Category

Grant-in-Aid for Scientific Research on Innovative Areas (Research in a proposed research area)

Allocation TypeSingle-year Grants
Review Section Complex systems
Research InstitutionAdvanced Telecommunications Research Institute International

Principal Investigator

石井 カルロス寿憲  株式会社国際電気通信基礎技術研究所, 石黒浩特別研究所, グループリーダー (30418529)

Project Period (FY) 2020-04-01 – 2022-03-31
Project Status Completed (Fiscal Year 2021)
Budget Amount *help
¥23,270,000 (Direct Cost: ¥17,900,000、Indirect Cost: ¥5,370,000)
Fiscal Year 2021: ¥11,700,000 (Direct Cost: ¥9,000,000、Indirect Cost: ¥2,700,000)
Fiscal Year 2020: ¥11,570,000 (Direct Cost: ¥8,900,000、Indirect Cost: ¥2,670,000)
Keywords音声情報処理 / 非言語情報処理 / パラ言語情報処理 / 人ロボットインタラクション / 音環境知能 / 社会的表出 / 動作生成 / アンドロイド / 人型ロボット / 対話状況認識
Outline of Research at the Start

従来のロボットやエージェントは、状況に合った適切な表出ができず、話し方もその動作も単調で人間らさが十分でない。本研究では、対話相手や状況に応じて人はどのように表出を変えるのかを表現できる「社会的表出」(Social Expression, SE)の数理モデルを明らかにし、それをロボットやエージェントとのインタラクションに実装する。インタラクションに関連する話し方および振る舞い方のパラメータを重視し、深層学習技術などに基づいたSEモデルの学習とその効果の検証を実施する。この研究により、人はロボットやエージェントとより自然に関われるようになり、ロボットやエージェントの利用範囲が格段に広がる。

Outline of Annual Research Achievements

本研究では、発話に伴う人間らしい自然な話し方と動作を持つ対話ロボット・エージェントの実現を目的とする。特に、対話相手や状況に応じて人はどのように表出を変えるのかを表現できる「社会的表出」の数理モデルを明らかにし、それをロボットやエージェントとのインタラクションに実装することを目指している。
初年度は、アンドロイドの丁寧な表出や怒りの表出について、人らしい振る舞いを実装し、ロボットが人を説得するタスクにおいて、どのような振る舞いが適しているのかについて評価実験を行ってきた。また、状況に合った視線制御のために3者対話に現れる視線および視線逸らしの理由の分析や、対話状況認識の観点から、深層学習による感情音声認識、ロボットに向けた暴言と冗談の識別などの研究にも取り組んできた。
今年度は、視線動作およびジェスチャ生成による個性の表出に取り組んだ。視線制御においては、眼球の動きも考慮し、3者対話データから参与役割を考慮した視線対象および視線逸らしの割合の分布および時間分布、視線を逸らした際の黒目の向きの分布を話者ごとに算出した。これらの分布をもとに、小型ロボットCommUに視線動作を実装し、被験者実験による印象評定を行った。外向性が異なる2人の話者のモデルによって生成した視線動作を評価した結果、同じ音声でも外向性の印象が変わることを確認した。
ジェスチャ生成においては、入力音声から抽出される韻律特徴を条件としたWGANによる手振りジェスチャを生成する深層学習モデルを構築し、人らしい自然な動作を生成できることを確認した。また、このモデルを拡張して、手の動きの大きさ・速さによって分類された3つのカテゴリーのラベルを条件に加えてモデルを再学習した。各カテゴリーを入力した際の動作を印象評定した結果、外向性の印象と相関した動作が生成できることをCGアバターおよび小型ロボットCommUにおいて確認した。

Research Progress Status

令和3年度が最終年度であるため、記入しない。

Strategy for Future Research Activity

令和3年度が最終年度であるため、記入しない。

Report

(2 results)
  • 2021 Annual Research Report
  • 2020 Annual Research Report
  • Research Products

    (26 results)

All 2021 2020

All Journal Article (14 results) (of which Peer Reviewed: 12 results,  Open Access: 2 results) Presentation (12 results) (of which Int'l Joint Research: 10 results)

  • [Journal Article] Prosodic and voice quality analyses of Japanese and Mandarin Chinese attitudinal speech: Japanese native speakers and Mandarin Chinese learners2021

    • Author(s)
      李キンゲツ・石井カルロス寿憲・林良子
    • Journal Title

      THE JOURNAL OF THE ACOUSTICAL SOCIETY OF JAPAN

      Volume: 77 Issue: 2 Pages: 112-119

    • DOI

      10.20697/jasj.77.2_112

    • NAID

      130007993002

    • ISSN
      0369-4232, 2432-2040
    • Year and Date
      2021-02-01
    • Related Report
      2021 Annual Research Report 2020 Annual Research Report
    • Peer Reviewed / Open Access
  • [Journal Article] Advocating Attitudinal Change Through Android Robot's Intention-Based Expressive Behaviors: Toward WHO COVID-19 Guidelines Adherence2021

    • Author(s)
      Ajibo Chinenye Augustine、Ishi Carlos Toshinori、Ishiguro Hiroshi
    • Journal Title

      IEEE Robotics and Automation Letters

      Volume: 6 Issue: 4 Pages: 6521-6528

    • DOI

      10.1109/lra.2021.3094783

    • Related Report
      2021 Annual Research Report
    • Peer Reviewed / Open Access
  • [Journal Article] Enabling Robots to Distinguish Between Aggressive and Joking Attitudes2021

    • Author(s)
      Maehama Kota、Even Jani、Ishi Carlos Toshinori、Kanda Takayuki
    • Journal Title

      IEEE Robotics and Automation Letters

      Volume: 6 Issue: 4 Pages: 8037-8044

    • DOI

      10.1109/lra.2021.3102974

    • Related Report
      2021 Annual Research Report
    • Peer Reviewed
  • [Journal Article] 3D skeletal movement-enhanced emotion recognition networks2021

    • Author(s)
      Shi Jiaqi、Liu Chaoran、Ishi Carlos Toshinori、Ishiguro Hiroshi
    • Journal Title

      APSIPA Transactions on Signal and Information Processing

      Volume: 10 Issue: 1 Pages: 1-12

    • DOI

      10.1017/atsip.2021.11

    • Related Report
      2021 Annual Research Report
    • Peer Reviewed
  • [Journal Article] Modeling the Conditional Distribution of Co-Speech Upper Body Gesture Jointly Using Conditional-GAN and Unrolled-GAN2021

    • Author(s)
      Wu Bowen、Liu Chaoran、Ishi Carlos Toshinori、Ishiguro Hiroshi
    • Journal Title

      Electronics

      Volume: 10 Issue: 3 Pages: 228-228

    • DOI

      10.3390/electronics10030228

    • Related Report
      2021 Annual Research Report 2020 Annual Research Report
    • Peer Reviewed
  • [Journal Article] MAEC: Multi-Instance Learning with an Adversarial Auto-Encoder-Based Classifier for Speech Emotion Recognition2021

    • Author(s)
      Fu Changzeng、Liu Chaoran、Ishi Carlos Toshinori、Ishiguro Hiroshi
    • Journal Title

      Proc. IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

      Volume: - Pages: 6299-6303

    • DOI

      10.1109/icassp39728.2021.9413640

    • Related Report
      2021 Annual Research Report
    • Peer Reviewed
  • [Journal Article] Analysis of Eye Gaze Reasons and Gaze Aversions During Three-Party Conversations2021

    • Author(s)
      Ishi Carlos Toshinori、Shintani Taiken
    • Journal Title

      Proc. Interspeech2021

      Volume: - Pages: 1972-1976

    • DOI

      10.21437/interspeech.2021-2134

    • Related Report
      2021 Annual Research Report
    • Peer Reviewed
  • [Journal Article] Probabilistic Human-like Gesture Synthesis from Speech using GRU-based WGAN2021

    • Author(s)
      Wu Bowen、Liu Chaoran、Ishi Carlos T.、Ishiguro Hiroshi
    • Journal Title

      Proc. of the 2021 International Conference on Multimodal Interaction (ICMI '21)

      Volume: - Pages: 194-201

    • DOI

      10.1145/3461615.3485407

    • Related Report
      2021 Annual Research Report
  • [Journal Article] Analysis of Role-Based Gaze Behaviors and Gaze Aversions, and Implementation of Robot’s Gaze Control for Multi-party Dialogue2021

    • Author(s)
      Shintani Taiken、Ishi Carlos T.、Ishiguro Hiroshi
    • Journal Title

      Proc. of the 9th International Conference on Human-Agent Interaction (HAI21)

      Volume: - Pages: 332-336

    • DOI

      10.1145/3472307.3484653

    • Related Report
      2021 Annual Research Report
  • [Journal Article] An End-to-end Multitask Learning Model to Improve Speech Emotion Recognition2021

    • Author(s)
      Fu Changzeng、Liu Chaoran、Ishi Carlos Toshinori、Ishiguro Hiroshi
    • Journal Title

      Proc. of 28th European Signal Processing Conference (EUSIPCO 2020)

      Volume: 1 Pages: 1-5

    • DOI

      10.23919/eusipco47968.2020.9287484

    • Related Report
      2020 Annual Research Report
    • Peer Reviewed
  • [Journal Article] Person-Directed Pointing Gestures and Inter-Personal Relationship: Expression of Politeness to Friendliness by Android Robots2020

    • Author(s)
      Ishi Carlos T.、Mikata Ryusuke、Ishiguro Hiroshi
    • Journal Title

      IEEE Robotics and Automation Letters

      Volume: 5 Issue: 4 Pages: 6081-6088

    • DOI

      10.1109/lra.2020.3011354

    • Related Report
      2020 Annual Research Report
    • Peer Reviewed
  • [Journal Article] Multi-Modality Emotion Recognition Model with GAT-Based Multi-Head Inter-Modality Attention2020

    • Author(s)
      Fu Changzeng、Liu Chaoran、Ishi Carlos Toshinori、Ishiguro Hiroshi
    • Journal Title

      Sensors

      Volume: 20 Issue: 17 Pages: 4894-4894

    • DOI

      10.3390/s20174894

    • Related Report
      2020 Annual Research Report
    • Peer Reviewed
  • [Journal Article] Analysis of body gestures in anger expression and evaluation in android robot2020

    • Author(s)
      Ajibo Chinenye Augustine、Ishi Carlos Toshinori、Mikata Ryusuke、Liu Chaoran、Ishiguro Hiroshi
    • Journal Title

      Advanced Robotics

      Volume: 34 Issue: 24 Pages: 1581-1590

    • DOI

      10.1080/01691864.2020.1855244

    • Related Report
      2020 Annual Research Report
    • Peer Reviewed
  • [Journal Article] Skeleton-Based Emotion Recognition Based on Two-Stream Self-Attention Enhanced Spatial-Temporal Graph Convolutional Network2020

    • Author(s)
      Shi Jiaqi、Liu Chaoran、Ishi Carlos Toshinori、Ishiguro Hiroshi
    • Journal Title

      Sensors

      Volume: 21 Issue: 1 Pages: 205-205

    • DOI

      10.3390/s21010205

    • Related Report
      2020 Annual Research Report
    • Peer Reviewed
  • [Presentation] MAEC: Multi-Instance Learning with an Adversarial Auto-Encoder-Based Classifier for Speech Emotion Recognition2021

    • Author(s)
      C. Fu, C. Liu, C. T. Ishi and H. Ishiguro
    • Organizer
      IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)
    • Related Report
      2021 Annual Research Report
    • Int'l Joint Research
  • [Presentation] Analysis of Eye Gaze Reasons and Gaze Aversions During Three-Party Conversations Recognition2021

    • Author(s)
      C. T. Ishi and T.Shintani
    • Organizer
      Interspeech 2021
    • Related Report
      2021 Annual Research Report
    • Int'l Joint Research
  • [Presentation] Probabilistic Human-like Gesture Synthesis from Speech using GRU-based WGAN2021

    • Author(s)
      B. Wu, C. Liu, C.T. Ishi, H. Ishiguro
    • Organizer
      International Conference on Multimodal Interaction (ICMI '21)
    • Related Report
      2021 Annual Research Report
    • Int'l Joint Research
  • [Presentation] Analysis of Role-Based Gaze Behaviors and Gaze Aversions, and Implementation of Robot's Gaze Control for Multi-party Dialogue2021

    • Author(s)
      T. Shintani, C.T. Ishi, H. Ishiguro
    • Organizer
      International Conference on Human-Agent Interaction (HAI21)
    • Related Report
      2021 Annual Research Report
    • Int'l Joint Research
  • [Presentation] Advocating Attitudinal Change Through Android Robot's Intention-Based Expressive Behaviors: Toward WHO COVID-19 Guidelines Adherence2021

    • Author(s)
      C. A. Ajibo, C.T. Ishi and H. Ishiguro
    • Organizer
      IROS2021
    • Related Report
      2021 Annual Research Report
    • Int'l Joint Research
  • [Presentation] Enabling Robots to Distinguish Between Aggressive and Joking Attitudes2021

    • Author(s)
      K. Maehama, J. Even, C.T. Ishi, T. Kanda
    • Organizer
      IROS2021
    • Related Report
      2021 Annual Research Report
    • Int'l Joint Research
  • [Presentation] 3D Skeletal Movement Enhanced Emotion Recognition Network2020

    • Author(s)
      J. Shi, C. Liu, C.T. Ishi, H. Ishiguro
    • Organizer
      Asia-Pacific Signal and Information Processing Association (Annual Summit and Conference 2020)
    • Related Report
      2020 Annual Research Report
    • Int'l Joint Research
  • [Presentation] Person-directed pointing gestures and inter-personal relationship: Expression of politeness to friendliness by android robots2020

    • Author(s)
      C.T. Ishi, R. Mikata, H. Ishiguro
    • Organizer
      International Conference on Intelligent Robots and Systems (IROS 2020)
    • Related Report
      2020 Annual Research Report
    • Int'l Joint Research
  • [Presentation] AAEC: An Adversarial Autoencoder-based Classifier for Audio Emotion Recognition2020

    • Author(s)
      C. Fu, J. Shi, C. Liu, C.T. Ishi, H. Ishiguro
    • Organizer
      MuSe 2020-The Multimodal Sentiment in Real-life Media Challenge (Conference: ACM Multimedia Conference 2020)
    • Related Report
      2020 Annual Research Report
    • Int'l Joint Research
  • [Presentation] An End-to-End Multitask Learning Model to Improve Speech Emotion Recognition2020

    • Author(s)
      C. Fu, C. Liu, C.T. Ishi and H. Ishiguro
    • Organizer
      28th European Signal Processing Conference (EUSIPCO 2020)
    • Related Report
      2020 Annual Research Report
    • Int'l Joint Research
  • [Presentation] Improving Conditional-GAN using Unrolled-GAN for the Generation of Co-speech Upper Body Gesture2020

    • Author(s)
      Bowen Wu, Chaoran Liu, Carlos Ishi, Hiroshi Ishiguro
    • Organizer
      第57回人工知能学会 AI チャレンジ研究会
    • Related Report
      2020 Annual Research Report
  • [Presentation] 複数人対話における役割に応じた視線のふるまいの解析とロボットへの実装2020

    • Author(s)
      新谷 太健, 石井カルロス寿憲, 石黒 浩
    • Organizer
      第57回人工知能学会 AI チャレンジ研究会
    • Related Report
      2020 Annual Research Report

URL: 

Published: 2020-04-28   Modified: 2023-12-25  

Information User Guide FAQ News Terms of Use Attribution of KAKENHI

Powered by NII kakenhi