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

Junpei Zhong*, Cornelius Weber, Stefan Wermter

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contribution

3 Citations (Scopus)

Abstract

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.

Original languageEnglish
Title of host publicationArtificial Neural Networks and Machine Learning, ICANN 2012 - 22nd International Conference on Artificial Neural Networks, Proceedings
Pages539-546
Number of pages8
Volume7552 LNCS
EditionPART 1
DOIs
Publication statusPublished - 2012
Externally publishedYes
Event22nd International Conference on Artificial Neural Networks, ICANN 2012 - Lausanne, Switzerland
Duration: 2012 Sept 112012 Sept 14

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
NumberPART 1
Volume7552 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

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

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Fingerprint

Dive into the research topics of 'Learning features and predictive transformation encoding based on a horizontal product model'. Together they form a unique fingerprint.

Cite this