Learning and recognition of multiple fluctuating temporal patterns using S-CTRNN

Shingo Murata, Hiroaki Arie, Tetsuya Ogata, Jun Tani, Shigeki Sugano

研究成果: Conference contribution

6 被引用数 (Scopus)

抄録

In the present study, we demonstrate the learning and recognition capabilities of our recently proposed recurrent neural network (RNN) model called stochastic continuous-time RNN (S-CTRNN). S-CTRNN can learn to predict not only the mean but also the variance of the next state of the learning targets. The network parameters consisting of weights, biases, and initial states of context neurons are optimized through maximum likelihood estimation (MLE) using the gradient descent method. First, we clarify the essential difference between the learning capabilities of conventional CTRNN and S-CTRNN by analyzing the results of a numerical experiment in which multiple fluctuating temporal patterns were used as training data, where the variance of the Gaussian noise varied among the patterns. Furthermore, we also show that the trained S-CTRNN can recognize given fluctuating patterns by inferring the initial states that can reproduce the patterns through the same MLE scheme as that used for network training.

本文言語English
ホスト出版物のタイトルArtificial Neural Networks and Machine Learning, ICANN 2014 - 24th International Conference on Artificial Neural Networks, Proceedings
出版社Springer Verlag
ページ9-16
ページ数8
ISBN(印刷版)9783319111780
DOI
出版ステータスPublished - 2014
イベント24th International Conference on Artificial Neural Networks, ICANN 2014 - Hamburg, Germany
継続期間: 2014 9 152014 9 19

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)
8681 LNCS
ISSN(印刷版)0302-9743
ISSN(電子版)1611-3349

Other

Other24th International Conference on Artificial Neural Networks, ICANN 2014
国/地域Germany
CityHamburg
Period14/9/1514/9/19

ASJC Scopus subject areas

  • 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(全般)

フィンガープリント

「Learning and recognition of multiple fluctuating temporal patterns using S-CTRNN」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