Transformation-Invariant Learning and Theoretical Guarantees for OOD Generalization
Keywords: pac learning guarantees, theory for distribution shifts, sample complexity, ood generalization, vc dimension
TL;DR: We initiate a theoretical study for learning under distribution shifts when the shifts are captured by a collection of data transformation maps.
Abstract: Learning with identical train and test distributions has been extensively investigated both practically and theoretically. Much remains to be understood, however, in statistical learning under distribution shifts. This paper focuses on a distribution shift setting where train and test distributions can be related by classes of (data) transformation maps. We initiate a theoretical study for this framework, investigating learning scenarios where the target class of transformations is either known or unknown. We establish learning rules and algorithmic reductions to Empirical Risk Minimization (ERM), accompanied with learning guarantees. We obtain upper bounds on the sample complexity in terms of the VC dimension of the class composing predictors with transformations, which we show in many cases is not much larger than the VC dimension of the class of predictors. We highlight that the learning rules we derive offer a game-theoretic viewpoint on distribution shift: a learner searching for predictors and an adversary searching for transformation maps to respectively minimize and maximize the worst-case loss.
Primary Area: Learning theory
Submission Number: 11887
Loading