Efficient Design of Spiking Neural Network With STDP Learning Based on Fast CORDIC
Efficient Design of Spiking Neural Network With STDP Learning Based on Fast CORDIC
复制标题
DOI：
10.1109/tcsi.2021.3061766
复制
发表时间：
2021-06-01
影响因子：
5.1
通讯作者：
Wang, Chao
中科院分区：
文献类型：
--
作者：
Wu, Jiajun;Zhan, Yi;Wang, Chao
在新兴的基于峰值神经网络（SNN）的神经形态硬件设计中，能源效率和在线学习是具有吸引力的优势，这主要得益于具有非线性动力学的生物启发局部学习，并以相关的硬件复杂性为代价。本文提出了一种采用快速坐标旋转数字计算机（CORDIC）算法的SNN设计，以实现高硬件效率的快速尖峰时间相关可塑性（STDP）学习。本研究提出了一种基于CORDIC的SNN系统设计和评估方法，从理论CORDIC级误差到应用级学习性能，寻找最优的CORDIC类型和精度。根据所提出的设计和评估方法，设计了一种基于快速收敛CORDIC的可重构SNN设计，以实现MNIST上的高分类精度、快速在线学习和良好的能效。利用SNN的容错和时分复用（TDM）策略，可重构SNN设计采用8位快速收敛CORDIC和基于TDM的硬件加速器来提高效率。FPGA实现结果证实，所提出的快速收敛CORDIC SNN设计在学习速度和能效方面比目前最先进的CORDIC方法高出38.5%-45.3%，在MNIST基准上，STDP学习速度为30.2 ns/SOP，能效为176.6 pJ/SOP，处理速度为6.1 ms/图像，在线学习收敛时间为21.4 s（平均达到最终精度的时间）。
In emerging Spiking Neural Network (SNN) based neuromorphic hardware design, energy efficiency and on-line learning are attractive advantages mainly contributed by bio-inspired local learning with nonlinear dynamics and at the cost of associated hardware complexity. This paper presents a novel SNN design employing fast COordinate Rotation DIgital Computer (CORDIC) algorithm to achieve fast spike timing-dependent plasticity (STDP) learning with high hardware efficiency. In this study, a system design and evaluation method of CORDIC-based SNN is proposed for finding optimal CORDIC type and precision, from theoretical CORDIC-level error to application-level learning performance. From the proposed design and evaluation method, a reconfigurable SNN design based on fast-convergence CORDIC is designed to achieve high classification accuracy on MNIST, fast on-line learning and good energy efficiency. By utilizing SNN's fault tolerance and time-division-multiplexing (TDM) strategy, the reconfigurable SNN design employs 8-bit fast-convergence CORDIC and TDM-based hardware accelerator for high efficiency. FPGA implementation results confirm that the proposed fast-convergence CORDIC SNN design outperforms the state-of-the-art CORDIC method by 38.5%-45.3% in terms of learning speed and energy efficiency, with the STDP learning of 30.2 ns/SOP, energy efficiency of 176.6 pJ/SOP, processing speed of 6.1 ms/image, and on-line learning convergence of 21.4 s (time to reach the final accuracy, on average), on MNIST benchmark.