Reducing ANN-SNN Conversion Error via Residual Membrane Potential Alignment

2026-08-14

AI Summary

This study addresses the accuracy degradation and high inference latency inherent in ANN-to-SNN conversion under small time steps by proposing a Residual Membrane Potential Alignment strategy. Through dynamic initial potential adjustment, RMPD regularization loss, and an SCR-Conv2d competitive refinement layer, this method effectively eliminates truncation bias while enhancing feature discriminability. Experiments on CIFAR and ImageNet datasets demonstrate significant improvements in conversion accuracy at t=2/4/8 with negligible computational overhead. Furthermore, the approach exhibits robust architectural generalizability, providing an effective solution for deploying low-latency, high-precision neuromorphic computing systems.

Abstract

Spiking Neural Networks (SNNs) serve as core architectures for neuromorphic computing thanks to event-driven operation and ultra-low power consumption. Direct SNN training is hindered by non-differentiable spikes that induce vanishing gradients and unstable optimization. ANN-SNN conversion circumvents such issues by reusing well-trained ANN weights for low-latency, energy-efficient inference. Nevertheless, existing conversion schemes suffer from severe accuracy drops at small timesteps, large inference delays and cumulative quantization errors, even with marginal performance loss at large $T$. To address these limitations, we first analyze flaws of conventional conversion pipelines from residual membrane potential statistics and propose a novel conversion strategy combining dynamic initial potential tuning and feature enhancement. We then introduce a regularization loss $\mathcal{L}_{\mathrm{RMPD}}$ to adapt initial potential of IF neurons and mitigate systematic truncation bias from boundary aggregation. A dedicated SCR-Conv2d competitive refinement layer with grouped convolution is further built to sharpen feature discrimination, eliminate redundant spikes and stabilize encoding under tiny time windows. Integrated with the state-of-the-art QCFS baseline, our approach delivers consistent low-latency performance gains and generalizes to ReLU CNNs, ANN Transformers, and multi-threshold SNN variants. Evaluations on CIFAR-10, CIFAR-100 and ImageNet verify prominent accuracy improvements at $T=2,4,8$, with negligible extra computation overhead. This work offers an effective conversion paradigm to facilitate real-world SNN deployment on neuromorphic chips.

Problem

• ANN-SNN conversion
• low latency
• conversion error
• residual membrane potential
• neuromorphic computing

Innovation

• Residual Membrane Potential Alignment
• ANN-SNN Conversion
• Regularization Loss L_RMPD
• SCR-Conv2d
• Low-Latency SNN