← Search

Vasco Portilheiro

2 accepted papers

2026

To Augment or Not to Augment? Diagnosing Distributional Symmetry Breaking

ICLR 2026poster

Symmetry-aware methods for machine learning, such as data augmentation and equivariant architectures, encourage correct model behavior on all transformations (e.g. rotations or permutations) of the original dataset. These methods can impart improved generalization and sample efficiency, under the as…

Cited by 0SourceScholar
2025

Improving Equivariant Networks with Probabilistic Symmetry Breaking

ICLR 2025poster

Equivariance encodes known symmetries into neural networks, often enhancing generalization. However, equivariant networks cannot *break* symmetries: the output of an equivariant network must, by definition, have at least the same self-symmetries as its input. This poses an important problem, both (1…

Cited by 8SourcePDFScholar