SpykeTorch: Efficient Simulation of Convolutional Spiking Neural Networks With at Most One Spike per Neuron
SpykeTorch: Efficient Simulation of Convolutional Spiking Neural Networks With at Most One Spike per Neuron
复制标题
DOI：
10.3389/fnins.2019.00625
复制
发表时间：
2019-07-12
影响因子：
4.3
通讯作者：
Masquelier, Timothee
中科院分区：
文献类型：
--
作者：
Mozafari, Milad;Ganjtabesh, Mohammad;Masquelier, Timothee
深度卷积尖峰神经网络（SNN）在人工智能（AI）任务中的应用最近引起了人们的极大兴趣，因为SNN是硬件友好且节能的。与非尖峰对应物不同，大多数现有的SNN仿真框架对于大规模AI任务来说实际上不够有效。在本文中，我们介绍了SpykeTorch，一个基于PyTorch的开源高速仿真框架。该框架模拟了卷积SNN，每个神经元最多有一个尖峰，并且采用了秩序编码方案。在学习规则方面，实现了尖峰时间依赖可塑性（STDP）和奖励调制STDP（R-STDP），但其他规则可以很容易地实现。除了上述属性，SpykeTorch是高度通用的，能够复制各种研究的结果。所提出的框架中的计算是基于张量的，完全由PyTorch函数完成，这反过来又带来了在CPU，GPU或多GPU平台上运行的即时优化能力。
Application of deep convolutional spiking neural networks (SNNs) to artificial intelligence (AI) tasks has recently gained a lot of interest since SNNs are hardware-friendly and energy-efficient. Unlike the non-spiking counterparts, most of the existing SNN simulation frameworks are not practically efficient enough for large-scale AI tasks. In this paper; we introduce SpykeTorch, an open-source high-speed simulation framework based on PyTorch. This framework simulates convolutional SNNs with at most one spike per neuron and the rank-order encoding scheme. In terms of learning rules, both spike-timing-dependent plasticity (STDP) and reward-modulated STDP (R-STDP) are implemented, but other rules could be implemented easily. Apart from the aforementioned properties, SpykeTorch is highly generic and capable of reproducing the results of various studies. Computations in the proposed framework are tensor-based and totally done by PyTorch functions, which in turn brings the ability of just-in-time optimization for running on CPUs, GPUs, or Multi-GPU platforms.