A Differentiable Point Process with Its Application to Spiking Neural Networks

TL;DR

This paper introduces a differentiable point process to enable low-variance path-wise gradient estimation for training spiking neural networks, improving upon previous high-variance score function methods.

Contribution

It develops a novel differentiable point process framework that allows for efficient gradient estimation in SNNs, addressing a key technical challenge.

Findings

The new gradient estimator reduces variance compared to score function methods.

Numerical simulations demonstrate improved training efficiency and accuracy.

The approach extends differentiability to arbitrary point process realizations.

Abstract

This paper is concerned about a learning algorithm for a probabilistic model of spiking neural networks (SNNs). Jimenez Rezende & Gerstner (2014) proposed a stochastic variational inference algorithm to train SNNs with hidden neurons. The algorithm updates the variational distribution using the score function gradient estimator, whose high variance often impedes the whole learning algorithm. This paper presents an alternative gradient estimator for SNNs based on the path-wise gradient estimator. The main technical difficulty is a lack of a general method to differentiate a realization of an arbitrary point process, which is necessary to derive the path-wise gradient estimator. We develop a differentiable point process, which is the technical highlight of this paper, and apply it to derive the path-wise gradient estimator for SNNs. We investigate the effectiveness of our gradient estimator through numerical simulation.