Attention and Extraction of Context with Formation of Adaptive Associative Memory through Reinforcement Learning

Summary A context-based attention task is employed in this paper. An Elman-type recurrent neural network is utilized to extract and keep the context information, and only the reinforcement signal that indicates whether the answer is correct or not is given. Through this learning, the function of an associative memory is observed in the Elman-type neural network. Adaptive formation of the basins are examined by varying the learning conditions.

Keywords: attention, associative memory, reinforcement learning, recurrent neural network, adaptive basin formation

Reference
4. Katsunari Shibata and Masanori Sugisaka:
Dynamics of a Recurrent Neural Network Acquired through the Learning of a Context-based Attention Task,
Proc. of AROB (Int'l Sympo. on Artificial Life and Robotics) 7th, pp. 152-155, 2002.1
pdf File (4 pages, 126kB)

3. Katsunari Shibata:
Formation of Attention and Associative Memory Based on Reinforcement Learning,
Proc. of ICCAS (Int'l Conf. on Control, Automation and Systems) 2001, pp. 9-12, 2001. 10
[Selective Attention, Associative Memory, Recurrent Neural Network, Reinforcement Learning]

pdf File (4 pages, 136kB)

2. 柴田克成, 杉坂政典:
報酬に基づく選択的注意の学習による文脈の抽出と連想,
平成13年度電気関係学会九州支部連合大会講演論文集, pp. 98, 2001. 10.
(in Japanese)

1. 柴田 克成, 伊藤 宏司:
認識の学習に基づく注意と連想記憶の形成,
Technical Report of IEICE (電子情報通信学会技術研究報告), NC99-137, pp. 153-160, 2000. 3.
(in Japanese)
pdf File (8 pages, 103kB)