Set Valued Predictions For Robust Domain Generalization
This work addresses the problem of domain generalization for machine learning models, offering a novel approach that could improve robustness in applications with diverse data distributions, though it appears incremental as it builds on existing DG frameworks.
The paper tackles the challenge of achieving robustness in Domain Generalization (DG) by proposing set-valued predictors instead of single-valued ones to enhance performance across unseen domains, and demonstrates its potential on real-world datasets from the WILDS benchmark.
Despite the impressive advancements in modern machine learning, achieving robustness in Domain Generalization (DG) tasks remains a significant challenge. In DG, models are expected to perform well on samples from unseen test distributions (also called domains), by learning from multiple related training distributions. Most existing approaches to this problem rely on single-valued predictions, which inherently limit their robustness. We argue that set-valued predictors could be leveraged to enhance robustness across unseen domains, while also taking into account that these sets should be as small as possible. We introduce a theoretical framework defining successful set prediction in the DG setting, focusing on meeting a predefined performance criterion across as many domains as possible, and provide theoretical insights into the conditions under which such domain generalization is achievable. We further propose a practical optimization method compatible with modern learning architectures, that balances robust performance on unseen domains with small prediction set sizes. We evaluate our approach on several real-world datasets from the WILDS benchmark, demonstrating its potential as a promising direction for robust domain generalization.