Understanding Domain Generalization: A Noise Robustness Perspective
This work addresses the problem of domain generalization for machine learning practitioners, but it is incremental as it reveals limitations in existing DG approaches without providing a new solution.
The paper investigates whether domain generalization algorithms outperform empirical risk minimization by analyzing label noise effects, finding that while DG algorithms show implicit robustness to label noise in synthetic settings, this does not lead to better performance on real-world benchmarks compared to ERM.
Despite the rapid development of machine learning algorithms for domain generalization (DG), there is no clear empirical evidence that the existing DG algorithms outperform the classic empirical risk minimization (ERM) across standard benchmarks. To better understand this phenomenon, we investigate whether there are benefits of DG algorithms over ERM through the lens of label noise. Specifically, our finite-sample analysis reveals that label noise exacerbates the effect of spurious correlations for ERM, undermining generalization. Conversely, we illustrate that DG algorithms exhibit implicit label-noise robustness during finite-sample training even when spurious correlation is present. Such desirable property helps mitigate spurious correlations and improve generalization in synthetic experiments. However, additional comprehensive experiments on real-world benchmark datasets indicate that label-noise robustness does not necessarily translate to better performance compared to ERM. We conjecture that the failure mode of ERM arising from spurious correlations may be less pronounced in practice.