PASCAL: Precise and Efficient ANN- SNN Conversion using Spike Accumulation and Adaptive Layerwise Activation

TL;DR

This paper introduces PASCAL, a method for converting ANNs to SNNs that maintains accuracy with minimal timesteps by leveraging mathematical equivalence and adaptive layerwise activation quantization.

Contribution

PASCAL provides a mathematically grounded ANN-SNN conversion method with layerwise quantization configuration, reducing inference timesteps while preserving accuracy.

Findings

Achieves approximately 74% accuracy on ImageNet with ResNet-34.

Reduces inference timesteps by 64 times compared to existing methods.

Maintains accuracy comparable to the original ANN.

Abstract

Spiking Neural Networks (SNNs) have been put forward as an energy-efficient alternative to Artificial Neural Networks (ANNs) since they perform sparse Accumulate operations instead of the power-hungry Multiply-and-Accumulate operations. ANN-SNN conversion is a widely used method to realize deep SNNs with accuracy comparable to that of ANNs.~\citeauthor{bu2023optimal} recently proposed the Quantization-Clip-Floor-Shift (QCFS) activation as an alternative to ReLU to minimize the accuracy loss during ANN-SNN conversion. Nevertheless, SNN inferencing requires a large number of timesteps to match the accuracy of the source ANN for real-world datasets. In this work, we propose PASCAL, which performs ANN-SNN conversion in such a way that the resulting SNN is mathematically equivalent to an ANN with QCFS-activation, thereby yielding similar accuracy as the source ANN with minimal inference timesteps. In addition, we propose a systematic method to configure the quantization step of QCFS activation in a layerwise manner, which effectively determines the optimal number of timesteps per layer for the converted SNN. Our results show that the ResNet-34 SNN obtained using PASCAL achieves an accuracy of $\approx$74\% on ImageNet with a 64$\times$ reduction in the number of inference timesteps compared to existing approaches.