6533b7d3fe1ef96bd12611ff

RESEARCH PRODUCT

On the Optimization of Self-Organizing Maps by Genetic Algorithms

subject

description

Publisher Summary This chapter reviews the research on the genetic optimization of self-organizing maps (SOMs). The optimization of learning rule parameters and of initial weights is able to improve network performance. The latter, however, requires chromosome sizes proportional to the size of the SOM and becomes unwieldy for large networks. The optimization of learning rule structures leads to self-organization processes of character similar to the standard learning rule. A particularly strong potential lies in the optimization of SOM topologies, which allows the study of global dynamical properties of SOMs and related models, as well as to develop tools for their analysis. Hierarchies of SOMs are sometimes used for classification tasks. A possible application of genetic algorithms (GAs) would be the evolution of those hierarchies as well as the filters used for data preprocessing. Finally, one of the most important open questions from the point of view of pure research as of applications is how network structures should be encoded in a genetic GA chromosome to attain a “creative” evolution process. It also mentions different approaches to code networks, most of them directed only to the creation of feed-forward networks. Perhaps in the future, together with GA, SMO will provide a valuable tool to study the important and intriguing question of modern research.