Practical Deep Learning with Bayesian Principles
This work addresses the challenge of making Bayesian deep learning practical for researchers and practitioners, though it is incremental in applying existing techniques to achieve this goal.
The paper tackles the impracticality of Bayesian methods in deep learning by demonstrating practical training with natural-gradient variational inference, achieving similar performance to Adam on large datasets like ImageNet while preserving benefits such as improved calibration and uncertainty handling.
Bayesian methods promise to fix many shortcomings of deep learning, but they are impractical and rarely match the performance of standard methods, let alone improve them. In this paper, we demonstrate practical training of deep networks with natural-gradient variational inference. By applying techniques such as batch normalisation, data augmentation, and distributed training, we achieve similar performance in about the same number of epochs as the Adam optimiser, even on large datasets such as ImageNet. Importantly, the benefits of Bayesian principles are preserved: predictive probabilities are well-calibrated, uncertainties on out-of-distribution data are improved, and continual-learning performance is boosted. This work enables practical deep learning while preserving benefits of Bayesian principles. A PyTorch implementation is available as a plug-and-play optimiser.