E(2) Equivariant Self-Attention for Radio Astronomy

TL;DR

This paper introduces group-equivariant self-attention models for radio galaxy classification, demonstrating improved performance, faster training, and better interpretability aligned with astronomers' feature attention.

Contribution

It presents novel group-equivariant self-attention architectures tailored for radio astronomy, enhancing explainability and efficiency over prior models.

Findings

Equivariant models require fewer training epochs.

Equivariant models outperform non-equivariant counterparts.

Models attend to features similar to human astronomers.

Abstract

In this work we introduce group-equivariant self-attention models to address the problem of explainable radio galaxy classification in astronomy. We evaluate various orders of both cyclic and dihedral equivariance, and show that including equivariance as a prior both reduces the number of epochs required to fit the data and results in improved performance. We highlight the benefits of equivariance when using self-attention as an explainable model and illustrate how equivariant models statistically attend the same features in their classifications as human astronomers.