Loss shaping enhances exact gradient learning with EventProp in Spiking Neural Networks

2022-12-02

Neuromorph. Comput. Eng.

AI Summary

Low-power keyword spotting on neuromorphic hardware: This work enhances the scalability and accuracy of EventProp for spiking neural networks (SNNs), addressing its limited loss-function support, high gradient estimation bias, and weak representational capacity. We propose three key innovations: (1) a novel loss-function shaping strategy to generalize EventProp to multi-class losses; (2) a temporally enhanced single-layer recurrent SNN architecture incorporating delay-line inputs, learnable heterogeneous time constants, and spike-latency encoding to improve temporal modeling; and (3) an efficient GPU-accelerated implementation in GeNN, integrating event-driven regularization and multi-class event-based data augmentation. Experiments demonstrate state-of-the-art performance on Spiking Heidelberg Digits and significant improvements over mainstream surrogate-gradient methods on Spiking Speech Commands—achieving 3× faster training and 4× lower memory consumption.

Abstract

Event-based machine learning promises more energy-efficient AI on future neuromorphic hardware. Here, we investigate how the recently discovered Eventprop algorithm for gradient descent on exact gradients in spiking neural networks can be scaled up to challenging keyword recognition benchmarks. We implemented Eventprop in the GPU-enhanced Neural Networks framework and used it for training recurrent spiking neural networks on the Spiking Heidelberg Digits and Spiking Speech Commands datasets. We found that learning depended strongly on the loss function and extended Eventprop to a wider class of loss functions to enable effective training. We then tested a large number of data augmentations and regularisations as well as exploring different network structures; and heterogeneous and trainable timescales. We found that when combined with two specific augmentations, the right regularisation and a delay line input, Eventprop networks with one recurrent layer achieved state-of-the-art performance on Spiking Heidelberg Digits and good accuracy on Spiking Speech Commands. In comparison to a leading surrogate-gradient-based SNN training method, our GeNN Eventprop implementation is 3X faster and uses 4X less memory. This work is a significant step towards a low-power neuromorphic alternative to current machine learning paradigms.

Problem

• Spiking Neural Networks
• Keyword Recognition
• Efficiency and Accuracy

Innovation

• Eventprop
• Spiking Neural Networks (SNN)
• GPU-accelerated Learning