脉冲神经网络研究进展综述

近年来,起源于计算神经科学的脉冲神经网络因其具有丰富的时空动力学特征、多样的编码机制、契合硬件的事件驱动特性等优势,在神经形态工程和类脑计算领域已得到广泛的关注.脉冲神经网络与当前计算机科学导向的以深度卷积网络为代表的人工神经网络的交叉融合被认为是发展人工通用智能的有力途径.对此,回顾了脉冲神经网络的发展历程,将其划分为神经元模型、训练算法、编程框架、数据集以及硬件芯片等5个重点方向,全方位介绍脉冲神经网络的最新进展和内涵,讨论并分析了脉冲神经网络领域各个重点方向的发展机遇和挑战.希望本综述能够吸引不同学科的研究者,通过跨学科的思想交流与合作研究,推动脉冲神经网络领域的发展.

Spiking neural networks:A survey on recent advances and new directions

Over the past few years, spiking neural networks(SNNs) originated from computational neuroscience have received extensive attention in the field of neuromorphic engineering and brain inspired computing due to their rich spatio-temporal dynamics, diverse coding schemes, and event-driven characteristics that naturally fit the neuromorphic hardware. Combining neuroscience-oriented spiking neural networks with computer-science-oriented artificial neural networks such as deep convolutional neural networks, is considered as a promising approach to artificial general intelligence. In this survey, we review the recent advances of spiking neural networks and focus on five major research areas, which we define as spiking neuron models, SNN algorithms, programming frameworks, datasets and neuromorphic hardware, and conclude with a broad discussion on the opportunities and challenges. The goals of this work are to provide an exhaustive review of spiking neural networks, attract researchers in different disciplines and motivate new works in this field through interdisciplinary exchange of ideas and collaborative research.