Domain Generalization In Robust Invariant Representation
This addresses the problem of domain generalization for robust invariant representations, which is incremental as it builds on existing unsupervised invariance methods.
The paper investigates whether model representations invariant to transformations in a seen domain remain invariant in unseen domains, finding that invariant models learn robust latent representations that handle distribution shifts effectively.
Unsupervised approaches for learning representations invariant to common transformations are used quite often for object recognition. Learning invariances makes models more robust and practical to use in real-world scenarios. Since data transformations that do not change the intrinsic properties of the object cause the majority of the complexity in recognition tasks, models that are invariant to these transformations help reduce the amount of training data required. This further increases the model's efficiency and simplifies training. In this paper, we investigate the generalization of invariant representations on out-of-distribution data and try to answer the question: Do model representations invariant to some transformations in a particular seen domain also remain invariant in previously unseen domains? Through extensive experiments, we demonstrate that the invariant model learns unstructured latent representations that are robust to distribution shifts, thus making invariance a desirable property for training in resource-constrained settings.