She Xueyuan, Dash Saurabh, Kim Daehyun, Mukhopadhyay Saibal
Department of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, GA, United States.
Front Neurosci. 2021 Jan 14;14:615756. doi: 10.3389/fnins.2020.615756. eCollection 2020.
This paper introduces a heterogeneous spiking neural network (H-SNN) as a novel, feedforward SNN structure capable of learning complex spatiotemporal patterns with spike-timing-dependent plasticity (STDP) based unsupervised training. Within H-SNN, hierarchical spatial and temporal patterns are constructed with convolution connections and memory pathways containing spiking neurons with different dynamics. We demonstrate analytically the formation of long and short term memory in H-SNN and distinct response functions of memory pathways. In simulation, the network is tested on visual input of moving objects to simultaneously predict for object class and motion dynamics. Results show that H-SNN achieves prediction accuracy on similar or higher level than supervised deep neural networks (DNN). Compared to SNN trained with back-propagation, H-SNN effectively utilizes STDP to learn spatiotemporal patterns that have better generalizability to unknown motion and/or object classes encountered during inference. In addition, the improved performance is achieved with 6x fewer parameters than complex DNNs, showing H-SNN as an efficient approach for applications with constrained computation resources.
本文介绍了一种异构脉冲神经网络（H-SNN），它是一种新颖的前馈脉冲神经网络结构，能够通过基于脉冲时间依赖可塑性（STDP）的无监督训练学习复杂的时空模式。在H-SNN中，分层的空间和时间模式通过卷积连接和包含具有不同动态特性的脉冲神经元的记忆路径构建而成。我们通过分析证明了H-SNN中长期和短期记忆的形成以及记忆路径的不同响应函数。在模拟中，该网络针对移动对象的视觉输入进行测试，以同时预测对象类别和运动动态。结果表明，H-SNN实现的预测准确率与有监督深度神经网络（DNN）相当或更高。与使用反向传播训练的脉冲神经网络相比，H-SNN有效利用STDP来学习时空模式，这些模式对推理过程中遇到的未知运动和/或对象类别具有更好的泛化能力。此外，与复杂的DNN相比，H-SNN以少6倍的参数实现了更好的性能，表明H-SNN是一种适用于计算资源受限应用的有效方法。
