How Lateral Interaction Develops in a Self-Organizing Feature Map (1993)
| Venue: | In Proceedings of the IEEE International Conference on Neural Networks |
| Citations: | 14 - 4 self |
BibTeX
@INPROCEEDINGS{Sirosh93howlateral,
author = {Joseph Sirosh and Risto Miikkulainen},
title = {How Lateral Interaction Develops in a Self-Organizing Feature Map},
booktitle = {In Proceedings of the IEEE International Conference on Neural Networks},
year = {1993},
pages = {1360--1365}
}
Bookmark
 |
 |
 |
 |
 |
 |
OpenURL
Abstract
A biologically motivated mechanism for self-organizing a neural network with modifiable lateral connections is presented. The weight modification rules are purely activity-dependent, unsupervised and local. The lateral interaction weights are initially random but develop into a "Mexican hat" shape around each neuron. At the same time, the external input weights self-organize to form a topological map of the input space. The algorithm demonstrates how self-organization can bootstrap itself using input information. Predictions of the algorithm agree very well with experimental observations on the development of lateral connections in cortical feature maps. I. Introduction Two-dimensional topological maps of sensory input are present in various cortices of the brain. They are believed to develop in a self-organizing process based on cooperation and competition between neurons [4, 13, 14]. The Self-Organizing Feature Map (SOFM) algorithm [5, 6] is a computational model of this process. T...
