Quantization-Aware In-situ Training for Reliable and Accurate Edge AI
De Lima JPC, Carro L (2022)
Publication Type: Conference contribution
Publication year: 2022
Publisher: Institute of Electrical and Electronics Engineers Inc.
Pages Range: 1497-1502
Conference Proceedings Title: Proceedings of the 2022 Design, Automation and Test in Europe Conference and Exhibition, DATE 2022
Event location: Virtual, Online, BEL
ISBN: 9783981926361
DOI: 10.23919/DATE54114.2022.9774657
Abstract
In-memory analog computation based on memristor crossbars has become the most promising approach for DNN inference. Because compute and memory requirements are larger during training, memristive crossbars are also an alternative to train DNN models within a feasible energy budget for edge devices, especially in the light of trends towards security, privacy, latency, and energy reduction, by avoiding data transfer over the Internet. To enable online training and inference on the same device, however, there are still challenges related to different minimum bitwidth needed in each phase, and memristor non-idealities to be addressed. We provide an in-situ training framework that allows the network to adapt to hardware imperfections, while practically eliminating errors from weight quantization. We validate our methodology with image classifiers, namely MNIST and CIFAR10, by training NN models with 8-bit weights and quantizing to 2 bits. The training algorithm recovers up to 12 % of the accuracy lost to quantization errors even under high variability, reduces training energy by up to 6 ×, and allows for energy-efficient inferences using a single cell per synapse, hence enhancing robustness and accuracy for a smooth training-to-inference transition.
Involved external institutions
Brazil (BR)How to cite
APA:
De Lima, J.P.C., & Carro, L. (2022). Quantization-Aware In-situ Training for Reliable and Accurate Edge AI. In Cristiana Bolchini, Ingrid Verbauwhede, Ioana Vatajelu (Eds.), Proceedings of the 2022 Design, Automation and Test in Europe Conference and Exhibition, DATE 2022 (pp. 1497-1502). Virtual, Online, BEL: Institute of Electrical and Electronics Engineers Inc..
MLA:
De Lima, Joao Paulo C., and Luigi Carro. "Quantization-Aware In-situ Training for Reliable and Accurate Edge AI." Proceedings of the 2022 Design, Automation and Test in Europe Conference and Exhibition, DATE 2022, Virtual, Online, BEL Ed. Cristiana Bolchini, Ingrid Verbauwhede, Ioana Vatajelu, Institute of Electrical and Electronics Engineers Inc., 2022. 1497-1502.
BibTeX: Download