• 研究課題をさがす
  • 研究者をさがす
  • KAKENの使い方
  1. 課題ページに戻る

2020 年度 実施状況報告書

Learning conversation agents for long-vision response generation in multi-round communications

研究課題

研究課題/領域番号 19K20345
研究機関徳島大学

研究代表者

康 シン  徳島大学, 大学院社会産業理工学研究部(理工学域), 助教 (80777350)

研究期間 (年度) 2019-04-01 – 2022-03-31
キーワードlong-vision conversation
研究実績の概要

The purpose of this research plan is to learn the long-vision conversation agents from scratch by allowing them to make sequential decisions in a multi-round communication. For these agents, the language ability, the self-concept identification ability, and the strategic conversation ability would be developed through three strategic communication games respectively.
In the second year, we collected the data of conversation for learning the self-concept identification ability for the conversation agents. The data was composed of two parts, which were the conversation turns collected from Twitter and the tags that should represent the meaning of the conversation. We employed a latent dirichlet allocation (LDA) model and a rule based method for the tag extraction. These tags were treated as the self-identities for the agents to learn.
The agents learned two neural networks which are a generator and a discriminator. The generator network was responsible to generate sequences of word tokens which represents its self-identity, that is, the tag given by the LDA model or the rule-based method, and the discriminator network was responsible to discriminate the identities of the other agents by analyzing the word tokens generated by those agents.

現在までの達成度 (区分)
現在までの達成度 (区分)

3: やや遅れている

理由

The experiment is based on an inter-conversation between agents, which requires them to generate the reasonable and meaningful responses for each conversation turn. Because generation of the agent responses were purely based on a generative model with a generative adversarial learning framework, controlling the quality of responses was extremely difficult. The quality of agents' inter-conversations has shown significant defect compared to the quality of human conversations.

今後の研究の推進方策

In the future work, the quality of inter-conversation should be improved, while the strategic conversation ability of conversation agents should be learned.
To improve the conversation quality, we would consider replacing the generator with GPT-based models, the retrieval-based models, and other slot-based approaches.
To learn the strategic conversation ability of conversation agents, we will design a long-vision reward for the agents through a reinforcement learning framework. Three agents will be given two types of social identities, and they must learn to identify the similarity of each other and to avoid to be recognized as the distinct one, by following the conversation rules. Agents get rewarded separately based on their decisions made through out the reinforcement learning framework.

次年度使用額が生じた理由

理由:文房具が3月に納品となり、支払いが完了していないため。
使用計画:文房具の支払が4月に完了する予定である。

  • 研究成果

    (15件)

すべて 2021 2020

すべて 雑誌論文 (5件) (うち査読あり 5件、 オープンアクセス 2件) 学会発表 (10件) (うち国際学会 7件)

  • [雑誌論文] The AI Development Through Transforming Tacit Knowledge to Explicit Knowledge of Nurses' Dialogue for Patients with Dementia2021

    • 著者名/発表者名
      Hirokazu Ito, Kazuyuki Matsumoto, Xin Kang, Tetsuya Tanioka, Yuko Yasuhara, Rozzano De Castro Locsin and Fuji Ren
    • 雑誌名

      International Journal of Advanced Intelligence (IJAI)

      巻: 12 ページ: 11-21

    • 査読あり
  • [雑誌論文] Active learning with complementary sampling for instructing class-biased multi-label text emotion classification2020

    • 著者名/発表者名
      Kang Xin、Shi Xuefeng、Wu Yunong、Ren Fuji
    • 雑誌名

      IEEE Transactions on Affective Computing

      巻: - ページ: -

    • DOI

      10.1109/TAFFC.2020.3038401

    • 査読あり
  • [雑誌論文] Vowel priority lip matching scheme and similarity evaluation model based on humanoid robot Ren-Xin2020

    • 著者名/発表者名
      Liu Zheng、Kang Xin、Nishide Shun、Ren Fuji
    • 雑誌名

      Journal of Ambient Intelligence and Humanized Computing

      巻: - ページ: 1~12

    • DOI

      10.1007/s12652-020-02175-9

    • 査読あり / オープンアクセス
  • [雑誌論文] DFF-ResNet: An insect pest recognition model based on residual networks2020

    • 著者名/発表者名
      Liu Wenjie、Wu Guoqing、Ren Fuji、Kang Xin
    • 雑誌名

      Big Data Mining and Analytics

      巻: 3 ページ: 300~310

    • DOI

      10.26599/BDMA.2020.9020021

    • 査読あり / オープンアクセス
  • [雑誌論文] Surface defect detection of steel strips based on classification priority YOLOv3-dense network2020

    • 著者名/発表者名
      Zhang Jiaqiao、Kang Xin、Ni Hongjun、Ren Fuji
    • 雑誌名

      Ironmaking & Steelmaking

      巻: - ページ: 1~12

    • DOI

      10.1080/03019233.2020.1816806

    • 査読あり
  • [学会発表] TUA1 at the NTCIR-15 DialEval-1 Task2020

    • 著者名/発表者名
      Xin Kang, Yunong Wu and Fuji Ren
    • 学会等名
      Proceedings of the 15th NTCIR Conference on Evaluation of Information Access Technologies
    • 国際学会
  • [学会発表] Gait recognition in real environment using gait energy image generated by Mask R-CNN2020

    • 著者名/発表者名
      Shota Inui, Fuji Ren, Shun Nishide and Xin Kang
    • 学会等名
      2020 IEEE International Conference on Mechatronics and Automation (ICMA)
    • 国際学会
  • [学会発表] Construction of emergency medical support system based on understanding dialect intention2020

    • 著者名/発表者名
      Kyoya Abe, Fuji Ren, Shun Nishide and Xin Kang
    • 学会等名
      The 15th International Conference on Natural Language Processing and Knowledge Engineering
    • 国際学会
  • [学会発表] A fusion model for multi-label emotion classification based on BERT and topic clustering2020

    • 著者名/発表者名
      Fei Ding, Xin Kang, Shun Nishide, Guan Zhijin and Fuji Ren
    • 学会等名
      In International Symposium on Artificial Intelligence and Robotics 2020
    • 国際学会
  • [学会発表] Speech emotion recognition based on data enhancement in time-frequency domain2020

    • 著者名/発表者名
      LI Qianqian, Fuji Ren, Shen Xiaoyan and Xin Kang
    • 学会等名
      International Symposium on Artificial Intelligence and Robotics 2020
    • 国際学会
  • [学会発表] Collaborative filtering recommendation algorithm based on bisecting K-means clustering2020

    • 著者名/発表者名
      Liu Jia, Xin Kang, Shun Nishide and Fuji Ren
    • 学会等名
      International Symposium on Artificial Intelligence and Robotics 2020
    • 国際学会
  • [学会発表] Examination to develop the artificial intelligence through transforming tacit knowledge of nurses' dialogue for patients with dementia towards explicit knowledge2020

    • 著者名/発表者名
      Hirokazu Ito, Kazuyuki Matsumoto, Xin Kang, Tetsuya Tanioka, Yuko Yasuhara, Rozzano De Castro Locsin and Fuji Ren
    • 学会等名
      The 15th International Conference on Natural Language Processing and Knowledge Engineering (NLP-KE20)
    • 国際学会
  • [学会発表] ロボット教師における特定分野のQAシステムの構築2020

    • 著者名/発表者名
      門田 尚之, 任 福継, 西出 俊, 康 シン
    • 学会等名
      情報処理学会全国大会
  • [学会発表] ロボット教師における特定人物の音声の自動生成2020

    • 著者名/発表者名
      寺尾 渉吾, 任 福継, 西出 俊, 康 シン
    • 学会等名
      情報処理学会全国大会
  • [学会発表] Tweet日本語感情コーパスに対するBERTの効果の検証2020

    • 著者名/発表者名
      池上 達也, 任 福継, 西出 俊, 康 シン
    • 学会等名
      情報処理学会全国大会

URL: 

公開日: 2021-12-27  

サービス概要 検索マニュアル よくある質問 お知らせ 利用規程 科研費による研究の帰属

Powered by NII kakenhi