Accelerating Deep Neural Networks with Phase-Change Memory Devices

Andrea Fasoli

Abstract: In this chapter, we discuss recent advances in the hardware acceleration of deep neural networks with analog memory devices. Analog memory offers enormous potential to speed up computation in deep learning. We study the use of Phase-Change Memory (PCM) as the resistive element in a crossbar array that allows the multiply-accumulate operation in deep neural networks to be performed in-memory. With this promise comes several challenges, including the impact of conductance drift on deep neural network accuracy. Here we introduce popular neural network architectures and explain how to accelerate inference using PCM arrays. We present a technique to compensate for conductance drift (“slope correction”) to allow in-memory computing with PCM during inference to reach software-equivalent deep learning baselines for a broad variety of important neural network workloads.