Fast and energy-efficient neuromorphic deep learning with first-spike times

TL;DR

This paper introduces a learning rule based on first-spike times for neuromorphic networks, enabling efficient error backpropagation, and demonstrates its implementation on neuromorphic hardware with analysis of platform-specific effects.

Contribution

It provides a rigorous derivation of a spike-time-based learning rule and shows how it can be used for error backpropagation in hierarchical neuromorphic networks.

Findings

Successful implementation on BrainScaleS-2 system demonstrating speed and energy efficiency

Effective error backpropagation using spike time coding

Analysis of performance under neuromorphic hardware distortions

Abstract

For a biological agent operating under environmental pressure, energy consumption and reaction times are of critical importance. Similarly, engineered systems are optimized for short time-to-solution and low energy-to-solution characteristics. At the level of neuronal implementation, this implies achieving the desired results with as few and as early spikes as possible. With time-to-first-spike coding both of these goals are inherently emerging features of learning. Here, we describe a rigorous derivation of a learning rule for such first-spike times in networks of leaky integrate-and-fire neurons, relying solely on input and output spike times, and show how this mechanism can implement error backpropagation in hierarchical spiking networks. Furthermore, we emulate our framework on the BrainScaleS-2 neuromorphic system and demonstrate its capability of harnessing the system's speed and energy characteristics. Finally, we examine how our approach generalizes to other neuromorphic platforms by studying how its performance is affected by typical distortive effects induced by neuromorphic substrates.