Stochastic resonance neurons in artificial neural networks

TL;DR

This paper introduces stochastic resonance neurons in neural networks, leveraging noise to enhance performance and robustness, and reducing the number of neurons needed for accurate results.

Contribution

It proposes a novel neural network architecture that incorporates stochastic resonance, demonstrating improved efficiency and noise robustness over traditional designs.

Findings

Significant reduction in neuron count for given accuracy

Enhanced robustness against noise impacts

Potential for more efficient optical neural networks

Abstract

Many modern applications of the artificial neural networks ensue large number of layers making traditional digital implementations increasingly complex. Optical neural networks offer parallel processing at high bandwidth, but have the challenge of noise accumulation. We propose here a new type of neural networks using stochastic resonances as an inherent part of the architecture and demonstrate a possibility of significant reduction of the required number of neurons for a given performance accuracy. We also show that such a neural network is more robust against the impact of noise.