Federated Learning of Spiking Neural Networks under Heterogeneous Temporal Resolutions

TL;DR

This paper introduces a federated learning framework for spiking neural networks that effectively handles heterogeneous temporal resolutions across edge devices, improving training accuracy in time-series tasks.

Contribution

It proposes novel adaptation methods for federated learning that address temporal resolution mismatches in SNNs, enabling effective collaborative training across diverse hardware constraints.

Findings

The framework significantly recovers accuracy lost due to temporal heterogeneity.

It enables local training at different temporal resolutions while maintaining global model compatibility.

Experimental results on SHD and DVS-Gesture datasets demonstrate improved robustness.

Abstract

Spiking neural networks (SNNs) are biologically inspired energy-efficient models that use sparse binary spike-based communication between neurons, making them attractive for resource-constrained edge devices. Federated learning enables such devices to train collaboratively without sharing raw data. In time-series applications, edge devices often collect data at different time resolutions due to hardware and energy constraints. This temporal heterogeneity poses a fundamental challenge for federated learning: parameters learned at one temporal resolution do not necessarily transfer directly to another, which might result in the naive federated averaging being ineffective. Targeting SNNs and, more broadly, deep networks with stateful neurons, we propose a federated learning framework that addresses this temporal resolution mismatch. We investigate how neuron parameters learned from data at different temporal resolutions and model aggregation should be integrated. We evaluate the proposed framework across two SNN-native benchmark datasets (SHD and DVS-Gesture) under a range of resolution heterogeneity scenarios. Our results show that the proposed adaptation methods can substantially recover accuracy lost due to temporal mismatch, hence enabling each client to train at their local temporal resolution while remaining compatible with the global model.