Networks with input gates for situation-dependent input selection in reinforcement learning

Junichi Murata, Masafumi Suzuki, Kotaro Hirasawa

研究成果: 会議への寄与タイプ論文

2 引用 (Scopus)

抄録

A method is proposed for situation-dependent input selection and learning acceleration in Q-learning. Q-values are expressed by an RBF network which has an input gate attached to each of its input channels in order to capture, by learning, situation-dependent relevance or usefulness of the input.

元の言語英語
ページ5-10
ページ数6
出版物ステータス出版済み - 1 1 2002
イベント2002 International Joint Conference on Neural Networks (IJCNN '02) - Honolulu, HI, 米国
継続期間: 5 12 20025 17 2002

その他

その他2002 International Joint Conference on Neural Networks (IJCNN '02)
米国
Honolulu, HI
期間5/12/025/17/02

Fingerprint

Radial basis function networks
Reinforcement learning

All Science Journal Classification (ASJC) codes

  • Software
  • Artificial Intelligence

これを引用

Murata, J., Suzuki, M., & Hirasawa, K. (2002). Networks with input gates for situation-dependent input selection in reinforcement learning. 5-10. 論文発表場所 2002 International Joint Conference on Neural Networks (IJCNN '02), Honolulu, HI, 米国.

Networks with input gates for situation-dependent input selection in reinforcement learning. / Murata, Junichi; Suzuki, Masafumi; Hirasawa, Kotaro.

2002. 5-10 論文発表場所 2002 International Joint Conference on Neural Networks (IJCNN '02), Honolulu, HI, 米国.

研究成果: 会議への寄与タイプ論文

Murata, J, Suzuki, M & Hirasawa, K 2002, 'Networks with input gates for situation-dependent input selection in reinforcement learning' 論文発表場所 2002 International Joint Conference on Neural Networks (IJCNN '02), Honolulu, HI, 米国, 5/12/02 - 5/17/02, pp. 5-10.
Murata J, Suzuki M, Hirasawa K. Networks with input gates for situation-dependent input selection in reinforcement learning. 2002. 論文発表場所 2002 International Joint Conference on Neural Networks (IJCNN '02), Honolulu, HI, 米国.
Murata, Junichi ; Suzuki, Masafumi ; Hirasawa, Kotaro. / Networks with input gates for situation-dependent input selection in reinforcement learning. 論文発表場所 2002 International Joint Conference on Neural Networks (IJCNN '02), Honolulu, HI, 米国.6 p.
@conference{1e7fed1bf2ad44f0af4df825eddf380b,
title = "Networks with input gates for situation-dependent input selection in reinforcement learning",
abstract = "A method is proposed for situation-dependent input selection and learning acceleration in Q-learning. Q-values are expressed by an RBF network which has an input gate attached to each of its input channels in order to capture, by learning, situation-dependent relevance or usefulness of the input.",
author = "Junichi Murata and Masafumi Suzuki and Kotaro Hirasawa",
year = "2002",
month = "1",
day = "1",
language = "English",
pages = "5--10",
note = "2002 International Joint Conference on Neural Networks (IJCNN '02) ; Conference date: 12-05-2002 Through 17-05-2002",

}

TY - CONF

T1 - Networks with input gates for situation-dependent input selection in reinforcement learning

AU - Murata, Junichi

AU - Suzuki, Masafumi

AU - Hirasawa, Kotaro

PY - 2002/1/1

Y1 - 2002/1/1

N2 - A method is proposed for situation-dependent input selection and learning acceleration in Q-learning. Q-values are expressed by an RBF network which has an input gate attached to each of its input channels in order to capture, by learning, situation-dependent relevance or usefulness of the input.

AB - A method is proposed for situation-dependent input selection and learning acceleration in Q-learning. Q-values are expressed by an RBF network which has an input gate attached to each of its input channels in order to capture, by learning, situation-dependent relevance or usefulness of the input.

UR - http://www.scopus.com/inward/record.url?scp=0036072715&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0036072715&partnerID=8YFLogxK

M3 - Paper

SP - 5

EP - 10

ER -