Title: Exploration of Input Patterns for Enhancing the Performance of Liquid State Machines

Abstract: Spiking Neural Networks (SNN) have gained increasing attention for its low power consumption. But training SNN is challenging. Liquid State Machine (LSM), as a major type of Reservoir computing, has been widely recognized for its low training cost among SNNs. The exploration of LSM topology for enhancing performance often requires hyper-parameter search, which is both resource-expensive and time-consuming. We explore the influence of input scale reduction on LSM instead. There are two main reasons for studying input reduction of LSM. One is that the input dimension of large images requires efficient processing. Another one is that input exploration is generally more economic than architecture search. To mitigate the difficulty in effectively dealing with huge input spaces of LSM, and to find that whether input reduction can enhance LSM performance, we explore several input patterns, namely fullscale, scanline, chessboard, and patch. Several datasets have been used to evaluate the performance of the proposed input patterns, including two spatio image datasets and one spatio-temporal image database. The experimental results show that the reduced input under chessboard pattern improves the accuracy by up to 5%, and reduces execution time by up to 50% with up to 75\% less input storage than the fullscale input pattern for LSM.