Dynamic Training of Liquid State Machines

TL;DR

This paper presents a method to optimize the training of Liquid State Machines by identifying effective weight ranges, using spike metrics and graph-based initializations to enhance SNN performance.

Contribution

It introduces a novel approach to optimize SNN training by selecting effective weight ranges and employing graph-based initializations for improved accuracy.

Findings

Spike metrics improve output accuracy

Barabasi-Albert graph yields best results

Optimized weights reduce output error

Abstract

Spiking Neural Networks (SNNs) emerged as a promising solution in the field of Artificial Neural Networks (ANNs), attracting the attention of researchers due to their ability to mimic the human brain and process complex information with remarkable speed and accuracy. This research aimed to optimise the training process of Liquid State Machines (LSMs), a recurrent architecture of SNNs, by identifying the most effective weight range to be assigned in SNN to achieve the least difference between desired and actual output. The experimental results showed that by using spike metrics and a range of weights, the desired output and the actual output of spiking neurons could be effectively optimised, leading to improved performance of SNNs. The results were tested and confirmed using three different weight initialisation approaches, with the best results obtained using the Barabasi-Albert random graph method.