Learning to Generalize across Domains on Single Test Samples

TL;DR

This paper introduces a meta-learning approach that enables models to adapt to unseen target domains on single test samples without additional training, improving domain generalization performance.

Contribution

It proposes a novel meta-learning framework that formulates adaptation as a variational Bayesian inference problem for single-sample domain generalization.

Findings

Model achieves comparable or better performance than state-of-the-art methods.

Adaptation requires only one feed-forward pass at test time.

The approach effectively mimics domain shifts during training.

Abstract

We strive to learn a model from a set of source domains that generalizes well to unseen target domains. The main challenge in such a domain generalization scenario is the unavailability of any target domain data during training, resulting in the learned model not being explicitly adapted to the unseen target domains. We propose learning to generalize across domains on single test samples. We leverage a meta-learning paradigm to learn our model to acquire the ability of adaptation with single samples at training time so as to further adapt itself to each single test sample at test time. We formulate the adaptation to the single test sample as a variational Bayesian inference problem, which incorporates the test sample as a conditional into the generation of model parameters. The adaptation to each test sample requires only one feed-forward computation at test time without any fine-tuning or self-supervised training on additional data from the unseen domains. Extensive ablation studies demonstrate that our model learns the ability to adapt models to each single sample by mimicking domain shifts during training. Further, our model achieves at least comparable -- and often better -- performance than state-of-the-art methods on multiple benchmarks for domain generalization.