Comparison of Consolidation Methods for Predictive Learning of Time Series

Ryoichi Nakajo*, Tetsuya Ogata

*この研究の対応する著者

研究成果

抄録

In environments where various tasks are sequentially given to deep neural networks (DNNs), training methods are needed that enable DNNs to learn the given tasks continuously. A DNN is typically trained by a single dataset, and continuous learning of subsequent datasets causes the problem of catastrophic forgetting. Previous studies have reported results for consolidation learning methods in recognition tasks and reinforcement learning problems. However, those methods were validated on only a few examples of predictive learning for time series. In this study, we applied elastic weight consolidation (EWC) and pseudo-rehearsal to the predictive learning of time series and compared their learning results. Evaluating the latent space after the consolidation learning revealed that the EWC method acquires properties of the pre-training and subsequent datasets with the same distribution, and the pseudo-rehearsal method distinguishes the properties and acquires them with different distributions.

本文言語English
ホスト出版物のタイトルAdvances and Trends in Artificial Intelligence. Artificial Intelligence Practices - 34th International Conference on Industrial, Engineering and Other Applications of Applied Intelligent Systems, IEA/AIE 2021, Proceedings
編集者Hamido Fujita, Ali Selamat, Jerry Chun-Wei Lin, Moonis Ali
出版社Springer Science and Business Media Deutschland GmbH
ページ113-120
ページ数8
ISBN(印刷版)9783030794569
DOI
出版ステータスPublished - 2021
イベント34th International Conference on Industrial, Engineering and Other Applications of Applied Intelligent Systems, IEA/AIE 2021 - Virtual, Online
継続期間: 2021 7 262021 7 29

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)
12798 LNAI
ISSN(印刷版)0302-9743
ISSN(電子版)1611-3349

Conference

Conference34th International Conference on Industrial, Engineering and Other Applications of Applied Intelligent Systems, IEA/AIE 2021
CityVirtual, Online
Period21/7/2621/7/29

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

「Comparison of Consolidation Methods for Predictive Learning of Time Series」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