NV-BNN: An Accurate Deep Convolutional Neural Network Based on Binary STT-MRAM for Adaptive AI Edge

Abstract

Binary STT-MRAM is a highly anticipated embedded non-volatile memory technology in advanced logic nodes <28 nm. How to enable its in-memory computing (IMC) capability is critical for enhancing AI Edge. Based on the soon-available STT-MRAM, we report the first binary deep convolutional neural network (NV-BNN) capable of both local and remote learning. Exploiting intrinsic cumulative switching probability, accurate online training of CIFAR-10 color images (similar to 90%) is realized using a relaxed endurance spec (switching <= 20 times) and hybrid digital/IMC design. For offline training, the accuracy loss due to imprecise weight placement can be mitigated using a rapid non-iterative training-with-noise and fine-tuning scheme.