Domain Generalization via Model-Agnostic Learning of Semantic Features
This addresses the problem of deploying machine learning models to real-world conditions with unknown domain shifts, representing an incremental advance in domain generalization methods.
The paper tackles domain generalization by training models on multi-domain source data to directly generalize to unseen target domains, achieving new state-of-the-art results on two object recognition benchmarks and consistent improvements on a medical image segmentation task.
Generalization capability to unseen domains is crucial for machine learning models when deploying to real-world conditions. We investigate the challenging problem of domain generalization, i.e., training a model on multi-domain source data such that it can directly generalize to target domains with unknown statistics. We adopt a model-agnostic learning paradigm with gradient-based meta-train and meta-test procedures to expose the optimization to domain shift. Further, we introduce two complementary losses which explicitly regularize the semantic structure of the feature space. Globally, we align a derived soft confusion matrix to preserve general knowledge about inter-class relationships. Locally, we promote domain-independent class-specific cohesion and separation of sample features with a metric-learning component. The effectiveness of our method is demonstrated with new state-of-the-art results on two common object recognition benchmarks. Our method also shows consistent improvement on a medical image segmentation task.