A Low-Power Hybrid-Precision Neuromorphic Processor With INT8 Inference and INT16 Online Learning in 40-nm CMOS
- 21 August 2023
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems I: Regular Papers
- Vol. 70 (10) , 4028-4039
Abstract
In this work, we present a neuromorphic processor for artificial intelligence of things (AIoT) applications featuring low-power consumption, a small footprint, STDP-based online learning, and the ability to adapt to multiple applications. Hybrid precision, i.e., INT8 for inference and INT16 for training, is suggested to achieve balanced accuracy and energy efficiency. A precision-configurable leaky integrate-and-fire(LIF) neuron unit and a unified memory architecture are designed to maximize datapath reuse. A dynamic pruning technique is proposed to exploit the temporal sparsity, yielding synaptic operations reduction by 3.68x in training and 1.63x in inference, respectively. The design is implemented and fabricated in a 40-nm CMOS process, with a core area of 0.87 mm $^2$ . It is measured to consume a minimal power of 680 $\mu$ W at 70 MHz under a 0.75 V power supply, corresponding to 9.9 pJ per synaptic operation. Evaluated with typical spatial, temporal, and spatiotemporal datasets (MNIST, MIT-BIH, and N-MNIST), the proposed design achieve energy efficiency comparable to the best-in-class solutions with handcrafted training and customized ASICs, while demonstrating improved versatility across multiple applications with balanced accuracy, power consumption, and model adaptability.Keywords
Funding Information
- National Natural Science Foundation of China (92164301, 62076066)
- Shanghai Municipal Science and Technology Major Project (2021SHZDZX0103)
- Shanghai Institute of Intelligent Electronics and Systems
This publication has 29 references indexed in Scilit:
- Learning IoT in Edge: Deep Learning for the Internet of Things with Edge ComputingIEEE Network, 2018
- Loihi: A Neuromorphic Manycore Processor with On-Chip LearningIEEE Micro, 2018
- Philosophy of the Spike: Rate-Based vs. Spike-Based Theories of the BrainFrontiers in Systems Neuroscience, 2015
- A million spiking-neuron integrated circuit with a scalable communication network and interfaceScience, 2014
- Neuronal Dynamics2014
- SpiNNaker: A 1-W 18-Core System-on-Chip for Massively-Parallel Neural Network SimulationIEEE Journal of Solid-State Circuits, 2013
- Spike Timing–Dependent Plasticity: A Hebbian Learning RuleAnnual Review of Neuroscience, 2008
- Which Model to Use for Cortical Spiking Neurons?IEEE Transactions on Neural Networks, 2004
- Simple model of spiking neuronsIEEE Transactions on Neural Networks, 2003
- The other side of the engram: experience-driven changes in neuronal intrinsic excitabilityNature Reviews Neuroscience, 2003