EPOC: A 28-nm 5.3 pJ/SOP Event-Driven Parallel Neuromorphic Hardware With Neuromodulation-Based Online Learning

• Published in: IEEE transactions on biomedical circuits and systems Vol. 19; no. 3; pp. 629 - 644
• Main Authors: Chen, Faquan, Tian, Qingyang, Xie, Lisheng, Zhou, Yifan, Wu, Ziren, Wu, Liangshun, Ying, Rendong, Wen, Fei, Liu, Peilin
• Format: Journal Article
• Language: English
• Published: United States IEEE 01.06.2025; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Accuracy; Adaptability; Algorithms; artificial neural network; Benchmarks; Biological neural networks; Complexity theory; Computational efficiency; Computer architecture; Distance learning; Energy efficiency; Firing pattern; Formulations; Hardware; Humans; Learning; Machine Learning; Memory management; Microprocessors; Neural networks; Neural Networks, Computer; Neuromodulation; Neuromorphic computing; Neuromorphic hardware; Neuromorphics; Neurons; on-chip learning; Parallel processing; parallelism; Sparsity; spike neural network; spike-based sparsity; Streaming; System-on-chip; Training
• ISSN: 1932-4545; 1940-9990; 1940-9990
• DOI: 10.1109/TBCAS.2024.3470520

Bio-inspired neuromorphic hardware with learning ability is highly promising to achieve human-like intelligence, particularly in terms of high energy efficiency and strong environmental adaptability. Though many customized prototypes have demonstrated learning ability, learning on neuromorphic hardware still lacks a bio-plausible and unified learning framework, and inherent spike-based sparsity and parallelism have not been fully exploited, which fundamentally limits their computational efficiency and scale. Therefore, we develop a unified, event-driven, and massively parallel multi-core neuromorphic online learning processor, namely EPOC. We present a neuromodulation-based neuromorphic online learning framework to unify various learning algorithms, and EPOC supports high-accuracy local/global supervised Spike Neural Network (SNN) learning with a low-memory-demand streaming single-sample learning strategy through different neuromodulator formulations. EPOC leverages a novel event-driven computation method that fully exploits spike-based sparsity throughout the forward-backward learning phases, and parallel multi-channel and multi-core computing architecture, bringing 9.9<inline-formula><tex-math notation="LaTeX">\times</tex-math></inline-formula> time efficiency improvement compared with the baseline architecture. We synthesize EPOC in a 28-nm CMOS process and perform extensive benchmarking. EPOC achieves state-of-the-art learning accuracy of 99.2%, 98.2%, and 94.3% on the MNIST, NMNIST, and DVS-Gesture benchmarks, respectively. Local-learning EPOC achieves 2.9<inline-formula><tex-math notation="LaTeX">\times</tex-math></inline-formula> time efficiency improvement compared with the global learning counterpart. EPOC operates at a typical clock frequency of 100 MHz, providing a peak 328 GOPS/51 GSOPS throughput and a 5.3 pJ/SOP energy efficiency.