Please use this identifier to cite or link to this item: https://doi.org/10.1016/j.bica.2013.06.005
Title: A connectionist model of data compression in memory
Authors: Iyer, L.R.
Ho, S.-B. 
Keywords: Data compression
MPEG compression
Prediction
Spatial memory
Issue Date: Oct-2013
Citation: Iyer, L.R., Ho, S.-B. (2013-10). A connectionist model of data compression in memory. Biologically Inspired Cognitive Architectures 6 : 58-66. ScholarBank@NUS Repository. https://doi.org/10.1016/j.bica.2013.06.005
Abstract: Mental imagery is an integral part of daily life, yet it has been poorly studied. In order to effectively imagine, and make future predictions, it is necessary to obtain an accurate picture of the world. If one were to remember events in complete detail, the memory requirements would be oppressive. Hence, it is necessary to compress the data. In psychology, memory compression has been very poorly studied. On the other hand, in computer technology, video and image compression has been thoroughly researched and standards such as MPEG, JPEG and GIF are largely in use today. We take inspiration from these techniques to form a connectionist framework of data compression. We then apply this framework to a problem in spatial cognition - given the motion of an object in a particular trajectory, its future motion should be predicted. An initial solution to this problem without compression was demonstrated earlier. In this paper, we demonstrate that there is a large memory reduction compared to the earlier system, and that larger simulations that were previously not viable can be run in this system. © 2013 Elsevier B.V. All rights reserved.
Source Title: Biologically Inspired Cognitive Architectures
URI: http://scholarbank.nus.edu.sg/handle/10635/111512
ISSN: 2212683X
DOI: 10.1016/j.bica.2013.06.005
Appears in Collections:Staff Publications

Show full item record
Files in This Item:
There are no files associated with this item.

SCOPUSTM   
Citations

2
checked on Nov 23, 2020

WEB OF SCIENCETM
Citations

1
checked on Nov 23, 2020

Page view(s)

83
checked on Nov 21, 2020

Google ScholarTM

Check

Altmetric


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.