Energy Constraints Improve Liquid State Machine Performance

TL;DR

Applying metabolic energy constraints to liquid state machines can enhance their accuracy and efficiency by modulating reservoir dynamics, leading to significant improvements in specific tasks like seizure detection.

Contribution

This paper introduces a novel approach of incorporating energy constraints into liquid state machines, demonstrating their positive impact on network performance and dynamics.

Findings

4.25% accuracy improvement on seizure detection

6.9% reduction in reservoir spiking activity

Energy constraints influence reservoir dynamics, improving task performance

Abstract

A model of metabolic energy constraints is applied to a liquid state machine in order to analyze its effects on network performance. It was found that, in certain combinations of energy constraints, a significant increase in testing accuracy emerged; an improvement of 4.25% was observed on a seizure detection task using a digital liquid state machine while reducing overall reservoir spiking activity by 6.9%. The accuracy improvements appear to be linked to the energy constraints' impact on the reservoir's dynamics, as measured through metrics such as the Lyapunov exponent and the separation of the reservoir.