Bayesian Neural Networks at Scale: A Performance Analysis and Pruning Study

TL;DR

This paper evaluates the scalability of Bayesian neural networks using high-performance computing, compares their training performance, and introduces a pruning method that reduces network size with minimal accuracy loss.

Contribution

It provides a detailed performance analysis of training BNNs at scale, compares them to conventional networks, and introduces BPrune for effective network pruning.

Findings

Pruning up to 80% of the network results in only 7% accuracy loss.

Distributed training enables scalable BNN training on high-performance clusters.

Pruning accelerates inference without significant accuracy degradation.

Abstract

Bayesian neural Networks (BNNs) are a promising method of obtaining statistical uncertainties for neural network predictions but with a higher computational overhead which can limit their practical usage. This work explores the use of high performance computing with distributed training to address the challenges of training BNNs at scale. We present a performance and scalability comparison of training the VGG-16 and Resnet-18 models on a Cray-XC40 cluster. We demonstrate that network pruning can speed up inference without accuracy loss and provide an open source software package, {\it{BPrune}} to automate this pruning. For certain models we find that pruning up to 80\% of the network results in only a 7.0\% loss in accuracy. With the development of new hardware accelerators for Deep Learning, BNNs are of considerable interest for benchmarking performance. This analysis of training a BNN at scale outlines the limitations and benefits compared to a conventional neural network.