Domain Generalization via Nuclear Norm Regularization
This addresses the challenge of generalizing to unseen domains for machine learning systems, but it is incremental as it builds on existing regularization methods.
The paper tackles the problem of domain generalization by proposing nuclear norm regularization to learn domain-invariant features, achieving improvements such as 1.7% and 0.9% test accuracy on the DomainBed benchmark.
The ability to generalize to unseen domains is crucial for machine learning systems deployed in the real world, especially when we only have data from limited training domains. In this paper, we propose a simple and effective regularization method based on the nuclear norm of the learned features for domain generalization. Intuitively, the proposed regularizer mitigates the impacts of environmental features and encourages learning domain-invariant features. Theoretically, we provide insights into why nuclear norm regularization is more effective compared to ERM and alternative regularization methods. Empirically, we conduct extensive experiments on both synthetic and real datasets. We show nuclear norm regularization achieves strong performance compared to baselines in a wide range of domain generalization tasks. Moreover, our regularizer is broadly applicable with various methods such as ERM and SWAD with consistently improved performance, e.g., 1.7% and 0.9% test accuracy improvements respectively on the DomainBed benchmark.