Future University Hakodate Academic Archive >
研究者 >
複雑系知能学科 >
斉藤 朝輝 >

このアイテムの引用には次の識別子を使用してください: http://hdl.handle.net/10445/6017

タイトル: Dynamical Singularities in Online Learning of Recurrent Neural Networks
著者: Saito, Asaki
Taiji, Makoto
Ikegami, Takashi
アブストラクト: We numerically and theoretically demonstrate various singularities, as a dynamical system, of a simple online learning system of a recurrent neural network (RNN) where RNN performs the one-step prediction of a time series generated by a one-dimensional map. More specifically, we show first through numerical simulations that the learning system exhibits singular behaviors (“neutral behaviors”) different from ordinary chaos, such as almost zero finite-time Lyapunov exponents, as well as inaccessibility and power-law decay of the distribution of learning times (transient times). Also, we show through linear stability analysis that, as a dynamical system, the learning system is represented by a singular map whose Jacobian matrix has eigenvalue unity in the whole phase space. In particular, we state that the singularity as a dynamical system (shown by the second method) provides a basic reason for the neutral behaviors (shown by the first method) exhibited by the learning system.
研究業績種別: 国際会議/International Conference
資料種別: Conference Paper
単著共著: 共著/joint
発表雑誌名,発表学会名など: Proceedings of the 2007 IEEE Symposium on Foundations of Computational Intelligence (FOCI 2007)
開始ページ: 174
終了ページ: 179
年月日: 2007年
出版社: IEEE
出現コレクション:斉藤 朝輝


ファイル 記述 サイズフォーマット
saito_2007_01_IEEES001P026.pdf3.36 MBAdobe PDF見る/開く



Powered by DSpace Software Copyright © 2002-2007 MIT and Hewlett-Packard