Pith. sign in

REVIEW 2 cited by

The Lie Derivative for Measuring Learned Equivariance

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2210.02984 v2 pith:IVYTY5M5 submitted 2022-10-06 cs.LG cs.AIcs.CVstat.ML

classification cs.LGcs.AIcs.CVstat.ML
keywords equivariancearchitecturederivativemodelmodelstrainingtransformersconvolutional
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Equivariance guarantees that a model's predictions capture key symmetries in data. When an image is translated or rotated, an equivariant model's representation of that image will translate or rotate accordingly. The success of convolutional neural networks has historically been tied to translation equivariance directly encoded in their architecture. The rising success of vision transformers, which have no explicit architectural bias towards equivariance, challenges this narrative and suggests that augmentations and training data might also play a significant role in their performance. In order to better understand the role of equivariance in recent vision models, we introduce the Lie derivative, a method for measuring equivariance with strong mathematical foundations and minimal hyperparameters. Using the Lie derivative, we study the equivariance properties of hundreds of pretrained models, spanning CNNs, transformers, and Mixer architectures. The scale of our analysis allows us to separate the impact of architecture from other factors like model size or training method. Surprisingly, we find that many violations of equivariance can be linked to spatial aliasing in ubiquitous network layers, such as pointwise non-linearities, and that as models get larger and more accurate they tend to display more equivariance, regardless of architecture. For example, transformers can be more equivariant than convolutional neural networks after training.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Equivariant spatio-hemispherical networks for diffusion MRI deconvolution

    eess.IV 2024-11 conditional novelty 6.0 of 10

    Exploiting the antipodal symmetry of diffusion MRI signals, the authors build an E(3)xSO(3)-equivariant deconvolution network on hemispheres, cutting computation 2 to 5 times while matching or improving fiber orientat...

  2. Enhancing Early Diabetic Retinopathy Detection through Synthetic DR1 Image Generation: A StyleGAN3 Approach

    eess.IV 2025-01 reject novelty 4.0 of 10

    A StyleGAN3 model generates realistic synthetic DR1 fundus images with good FID/KID scores, but the paper does not test whether these images improve any diabetic retinopathy classifier.

Pith tools