Learning features and predictive transformation encoding based on a horizontal product model

Junpei Zhong, Cornelius Weber, Stefan Wermter

研究成果: Conference contribution

3 被引用数 (Scopus)

抄録

The visual system processes the features and movement of an object in separate pathways, called the ventral and dorsal streams. To integrate this principle in a functional model, a recurrent predictive network with a horizontal product is introduced. Learned in an unsupervised manner, two sets of hidden units represent cells in the ventral and dorsal pathways, respectively. Experiments show that the activity in the ventral-like units persists, given that the same feature appears in the receptive field, whilst the activity in the dorsal-like units shows a fluctuating pattern with different directions of object movements. Moreover, we show that the position information predicts the input's future position taking into account its moving direction due to the direction-selective responses of the dorsal-like units.

本文言語English
ホスト出版物のタイトルArtificial Neural Networks and Machine Learning, ICANN 2012 - 22nd International Conference on Artificial Neural Networks, Proceedings
ページ539-546
ページ数8
7552 LNCS
PART 1
DOI
出版ステータスPublished - 2012
外部発表はい
イベント22nd International Conference on Artificial Neural Networks, ICANN 2012 - Lausanne, Switzerland
継続期間: 2012 9 112012 9 14

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)
番号PART 1
7552 LNCS
ISSN(印刷版)03029743
ISSN(電子版)16113349

Other

Other22nd International Conference on Artificial Neural Networks, ICANN 2012
CountrySwitzerland
CityLausanne
Period12/9/1112/9/14

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

フィンガープリント 「Learning features and predictive transformation encoding based on a horizontal product model」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