On the Validity of Bayesian Neural Networks for Uncertainty Estimation

Abstract

Deep neural networks (DNN) are versatile parametric models utilised successfully in a diverse number of tasks and domains. However, they have limitations---particularly from their lack of robustness and over-sensitivity to out of distribution samples. Bayesian Neural Networks, due to their formulation under the Bayesian framework, provide a principled approach to building neural networks that address these limitations. This paper describes a study that empirically evaluates and compares Bayesian Neural Networks to their equivalent point estimate Deep Neural Networks to quantify the predictive uncertainty induced by their parameters, as well as their performance in view of this uncertainty. In this study, we evaluated and compared three point estimate deep neural networks against comparable Bayesian neural network alternatives using two well-known benchmark image classification datasets (CIFAR-10 and SVHN).

arXiv commentAICS2019, fixed typos, figures, tables

References in corpus (10)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Wide Residual Networks
• Averaging Weights Leads to Wider Optima and Better Generalization
• Identity Mappings in Deep Residual Networks
• A Simple Baseline for Bayesian Uncertainty in Deep Learning
• WAIC, but Why? Generative Ensembles for Robust Anomaly Detection
• Training Confidence-calibrated Classifiers for Detecting Out-of-Distribution Samples
• Stochastic Gradient Descent as Approximate Bayesian Inference
• Verified Uncertainty Calibration
• A Less Biased Evaluation of Out-of-distribution Sample Detectors

Cited by in corpus (5)

• Hands-on Bayesian Neural Networks -- a Tutorial for Deep Learning Users
• Explainable Artificial Intelligence for Bayesian Neural Networks: Towards trustworthy predictions of ocean dynamics
• Probabilistic Neural Data Fusion for Learning from an Arbitrary Number of Multi-fidelity Data Sets
• The missing radial velocities of Gaia: blind predictions for DR3
• The Ridgelet Prior: A Covariance Function Approach to Prior Specification for Bayesian Neural Networks