robust reservoir generation by correlation-based learning健壮的水库一代correlation-based学习.pdf

Hindawi Publishing Corporation Advances in Artificial Neural Systems Volume 2009, Article ID 467128, 7 pages doi:10.1155/2009/467128 Research Article Robust Reservoir Generation by Correlation-Based Learning Tadashi Yamazaki1 and Shigeru Tanaka2 1 Laboratory for Motor Learning Control, RIKEN Brain Science Institute, 2-1 Hirosawa, Wako, Saitama 351-0198, Japan 2 Laboratory for Visual Neurocomputing, RIKEN Brain Science Institute, 2-1 Hirosawa, Wako, Saitama 351-0198, Japan Correspondence should be addressed to Shigeru Tanaka, shigeru@riken.jp Received 12 March 2008; Revised 7 August 2008; Accepted 9 September 2008 Recommended by Akira Imada Reservoir computing (RC) is a new framework for neural computation. A reservoir is usually a recurrent neural network with fixed random connections. In this article, we propose an RC model in which the connections in the reservoir are modifiable. Specifically, we consider correlation-based learning (CBL), which modifies the connection weight between a given pair of neurons according to the correlation in their activities. We demonstrate that CBL enables the reservoir to reproduce almost the same spatiotemporal activity patterns in response to an identical input stimulus in the presence of noise. This result suggests that CBL enhances the robustness in the generation of the spatiotemporal activity pattern against noise in input signals. We apply our RC model to trace eyeblink conditioning. The reservoir bridged the gap of an interstimulus interval between the conditioned and unconditioned stimuli, and a readout neuron was able to learn and express the timed conditioned response. Copyright © 2009 T. Yamazaki and S. Tanaka. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use,