Deakin University
Browse

File(s) under permanent embargo

Extensions to knowledge acquisition and effect of multimodal representation in unsupervised learning

chapter
posted on 2009-01-01, 00:00 authored by Daswin De Silva, D Alahakoon, S Dharmage
The phenomenal behaviour and composition of human cognition is yet to be defined comprehensibly. Developing the same, artificially, is a foremost research area in artificial intelligence and related fields. In this chapter we look at advances made in the unsupervised learning paradigm (self organising methods) and its potential in realising artificial cognitive machines. The first section delineates intricacies of the process of learning in humans with an articulate discussion of the function of thought and the function of memory. The self organising method and the biological rationalisations that led to its development are explored in the second section. The next focus is the effect of structure restrictions on unsupervised learning and the enhancements resulting from a structure adapting learning algorithm. Generation of a hierarchy of knowledge using this algorithm will also be discussed. Section four looks at new means of knowledge acquisition through this adaptive unsupervised learning algorithm while the fifth examines the contribution of multimodal representation of inputs to unsupervised learning. The chapter concludes with a summary of the extensions outlined.

History

Title of book

Foundations of computational intelligence Volume 1 : learning and approximation

Series

Studies in computational intelligence ; v. 201

Chapter number

11

Pagination

281 - 305

Publisher

Springer

Place of publication

Berlin, Germany

ISBN-13

9783642010811

ISBN-10

3642010822

Language

eng

Publication classification

B1.1 Book chapter

Copyright notice

2009, Springer

Extent

15

Editor/Contributor(s)

A Hassanien, A Abraham, A Vasilakos, W Pedrycz

Usage metrics

    Research Publications

    Categories

    No categories selected

    Keywords

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC