We propose a technology-independent method, referred to as adjacent connection matrix (ACM), to efficiently map signed weight matrices to non-negative crossbar arrays. When compared to same-hardware-overhead mapping methods, using ACM leads to improvements of up to 20% in training accuracy for ResNet-20 with the CIFAR-10 dataset when training with 5-bit precision crossbar arrays or lower. When compared with strategies that use two elements to represent a weight, ACM achieves comparable training accuracies, while also offering area and read energy reductions of 2.3x and 7x, respectively. ACM also has a mild regularization effect that improves inference accuracy in crossbar arrays without any retraining or costly device/variation-aware training.
@article{arxiv.2004.06094,
title = {A Device Non-Ideality Resilient Approach for Mapping Neural Networks to Crossbar Arrays},
author = {Arman Kazemi and Cristobal Alessandri and Alan C. Seabaugh and X. Sharon Hu and Michael Niemier and Siddharth Joshi},
journal= {arXiv preprint arXiv:2004.06094},
year = {2020}
}