A Bit More Bayesian: Domain-Invariant Learning with Uncertainty
It improves domain generalization for cross-domain visual recognition, though it is incremental as it builds on existing Bayesian and invariance methods.
The paper tackles domain generalization by addressing domain shift and uncertainty using a probabilistic framework based on variational Bayesian inference, achieving state-of-the-art mean accuracy on four cross-domain visual recognition benchmarks.
Domain generalization is challenging due to the domain shift and the uncertainty caused by the inaccessibility of target domain data. In this paper, we address both challenges with a probabilistic framework based on variational Bayesian inference, by incorporating uncertainty into neural network weights. We couple domain invariance in a probabilistic formula with the variational Bayesian inference. This enables us to explore domain-invariant learning in a principled way. Specifically, we derive domain-invariant representations and classifiers, which are jointly established in a two-layer Bayesian neural network. We empirically demonstrate the effectiveness of our proposal on four widely used cross-domain visual recognition benchmarks. Ablation studies validate the synergistic benefits of our Bayesian treatment when jointly learning domain-invariant representations and classifiers for domain generalization. Further, our method consistently delivers state-of-the-art mean accuracy on all benchmarks.