<br>用于实现高精度和低延迟深度尖峰神经网络的通用 ANN 到 SNN 框架,Neural Networks

当前位置： X-MOL 学术 › Neural Netw. › 论文详情

Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)

用于实现高精度和低延迟深度尖峰神经网络的通用 ANN 到 SNN 框架
Neural Networks ( IF 6.0 ) Pub Date : 2024-03-15 , DOI: 10.1016/j.neunet.2024.106244
Yuchen Wang ₁ , Hanwen Liu ₁ , Malu Zhang ₁ , Xiaoling Luo ₂ , Hong Qu ₁

Affiliation

尖峰神经网络 (SNN) 因其生物学合理性、低功耗以及神经形态硬件实现的潜力而成为最著名的下一代计算模型之一。在获取可用 SNN 的各种方法中，将人工神经网络 (ANN) 转换为 SNN 是最具成本效益的方法。 ANN 到 SNN 转换工作的早期挑战围绕着转换后的 SNN 对转换错误的敏感性。最近的一些努力试图通过改变原始的人工神经网络来减轻这些转换错误。尽管它们能够提高 SNN 的准确性，但这些方法缺乏通用性，不能直接应用于大多数现有 ANN 的转换。在本文中，我们提出了一种名为 DNISNM 的框架，用于将 ANN 转换为 SNN，旨在解决由于 ANN 和 SNN 之间网络传输的离散性和异步性差异而产生的转换错误。 DNISNM 由两种机制组成，即基于数据的神经元初始化（DNI）和带记忆的签名神经元（SNM），旨在分别解决离散性和异步差异引起的错误。该框架不需要对原始 ANN 进行额外修改，可以提高 SNN 的精度性能，同时确保通用性、高精度和低推理延迟。我们在具有挑战性的对象识别数据集（包括 CIFAR10、CIFAR100 和 ImageNet-1k）上进行了实验验证。实验结果表明，即使在极低的延迟下，我们的框架转换的 SNN 也具有非常高的准确性。

"点击查看英文标题和摘要"

A universal ANN-to-SNN framework for achieving high accuracy and low latency deep Spiking Neural Networks

Spiking Neural Networks (SNNs) have become one of the most prominent next-generation computational models owing to their biological plausibility, low power consumption, and the potential for neuromorphic hardware implementation. Among the various methods for obtaining available SNNs, converting Artificial Neural Networks (ANNs) into SNNs is the most cost-effective approach. The early challenges in ANN-to-SNN conversion work revolved around the susceptibility of converted SNNs to conversion errors. Some recent endeavors have attempted to mitigate these conversion errors by altering the original ANNs. Despite their ability to enhance the accuracy of SNNs, these methods lack generality and cannot be directly applied to convert the majority of existing ANNs. In this paper, we present a framework named DNISNM for converting ANN to SNN, with the aim of addressing conversion errors arising from differences in the discreteness and asynchrony of network transmission between ANN and SNN. The DNISNM consists of two mechanisms, Data-based Neuronal Initialization (DNI) and Signed Neuron with Memory (SNM), designed to respectively address errors stemming from discreteness and asynchrony disparities. This framework requires no additional modifications to the original ANN and can result in SNNs with improved accuracy performance, simultaneously ensuring universality, high precision, and low inference latency. We verify it experimentally on challenging object recognition datasets, including CIFAR10, CIFAR100, and ImageNet-1k. Experimental results show that the SNN converted by our framework has very high accuracy even at extremely low latency.

更新日期：2024-03-15

点击分享查看原文

点击收藏

阅读更多本刊新发论文本刊介绍/投稿指南