Bayesian Variational Autoencoders for Unsupervised Out-of-Distribution Detection
This addresses AI safety concerns by improving OoD detection, but it appears incremental as it builds on existing variational autoencoder methods.
The paper tackles the problem of deep neural networks making unreliable predictions on out-of-distribution (OoD) test data by proposing a Bayesian variational autoencoder model that estimates a full posterior distribution over decoder parameters using stochastic gradient Markov chain Monte Carlo, and it demonstrates effectiveness through empirical results.
Despite their successes, deep neural networks may make unreliable predictions when faced with test data drawn from a distribution different to that of the training data, constituting a major problem for AI safety. While this has recently motivated the development of methods to detect such out-of-distribution (OoD) inputs, a robust solution is still lacking. We propose a new probabilistic, unsupervised approach to this problem based on a Bayesian variational autoencoder model, which estimates a full posterior distribution over the decoder parameters using stochastic gradient Markov chain Monte Carlo, instead of fitting a point estimate. We describe how information-theoretic measures based on this posterior can then be used to detect OoD inputs both in input space and in the model's latent space. We empirically demonstrate the effectiveness of our proposed approach.