Equi Tuning Group Equivariant Fine Tuning of Pretrained Models

Properties
authors	Sourya Basu

Main idea

Given non-equivariant pre-trained model \(M(x)\), define equivariant model \(M_G(x)\), as the average of the inverted predictions for all group actions on input \(x\)

\[ M_G(x) = \frac{1}{|G|} \sum_{g \in G} g^{-1} M(g x) \]

Abstract