Bayesian Inference on Binary Spiking Networks Leveraging Nanoscale Device Stochasticity

Abstract

Bayesian Neural Networks (BNNs) can overcome the problem of overconfidence that plagues traditional frequentist deep neural networks, and are hence considered to be a key enabler for reliable AI systems. However, conventional hardware realizations of BNNs are resource intensive, requiring the implementation of random number generators for synaptic sampling. Owing to their inherent stochasticity during programming and read operations, nanoscale memristive devices can be directly leveraged for sampling, without the need for additional hardware resources. In this paper, we introduce a novel Phase Change Memory (PCM)-based hardware implementation for BNNs with binary synapses. The proposed architecture consists of separate weight and noise planes, in which PCM cells are configured and operated to represent the nominal values of weights and to generate the required noise for sampling, respectively. Using experimentally observed PCM noise characteristics, for the exemplary Breast Cancer Dataset classification problem, we obtain hardware accuracy and expected calibration error matching that of an 8-bit fixed-point (FxP8) implementation, with projected savings of over 9$\times$ in terms of core area transistor count.

arXiv commentSubmitted and Accepted in ISCAS 2023

References in corpus (10)

• Categorical Reparameterization with Gumbel-Softmax
• On Calibration of Modern Neural Networks
• Accurate deep neural network inference using computational phase-change memory
• Markov Chain Monte Carlo and Variational Inference: Bridging the Gap
• Synaptic Plasticity Dynamics for Deep Continuous Local Learning (DECOLLE)
• An Introduction to Probabilistic Spiking Neural Networks: Probabilistic Models, Learning Rules, and Applications
• Committee machines -- a universal method to deal with non-idealities in memristor-based neural networks
• The Case for Bayesian Deep Learning
• An Overview of In-memory Processing with Emerging Non-volatile Memory for Data-intensive Applications
• All-Spin Bayesian Neural Networks

Cited by in corpus (1)

• Bayes2IMC: In-Memory Computing for Bayesian Binary Neural Networks