Robust population coding in free-energy-based reinforcement learning

Makoto Otsuka, Junichiro Yoshimoto, Kenji Doya

研究成果: 書籍/レポート タイプへの寄稿会議への寄与

1 被引用数 (Scopus)

抄録

We investigate the properties of free-energy-based reinforcement learning using a new experimental platform called the digit floor task. The simulation results showed the robustness of the reinforcement learning method against noise applied in both the training and testing phases. In addition, reward-dependent and reward-invariant representations were found in the distributed activation patterns of hidden units. The representations coded in a distributed fashion persisted even when the number of hidden nodes were varied.

本文言語英語
ホスト出版物のタイトルArtificial Neural Networks - ICANN 2008 - 18th International Conference, Proceedings
ページ377-386
ページ数10
PART 1
DOI
出版ステータス出版済み - 2008
外部発表はい
イベント18th International Conference on Artificial Neural Networks, ICANN 2008 - Prague, チェコ共和国
継続期間: 03-09-200806-09-2008

出版物シリーズ

名前Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
番号PART 1
5163 LNCS
ISSN(印刷版)0302-9743
ISSN(電子版)1611-3349

会議

会議18th International Conference on Artificial Neural Networks, ICANN 2008
国/地域チェコ共和国
CityPrague
Period03-09-0806-09-08

All Science Journal Classification (ASJC) codes

  • 理論的コンピュータサイエンス
  • コンピュータサイエンス一般

フィンガープリント

「Robust population coding in free-energy-based reinforcement learning」の研究トピックを掘り下げます。これらがまとまってユニークなフィンガープリントを構成します。

引用スタイル