Learning Layer-wise Equivariances Automatically using Gradients

Convolutions encode equivariance symmetries into neural networks leading to better generalisation performance. However, symmetries provide fixed hard constraints on the functions a network can represent, need to be specified in advance, and can not be adapted. Our goal is to allow flexible symmetry constraints that can automatically be learned from data using gradients. Learning symmetry and associated weight connectivity structures from scratch is difficult for two reasons. First, it requires efficient and flexible parameterisations of layer-wise equivariances. Secondly, symmetries act as constraints and are therefore not encouraged by training losses measuring data fit. To overcome these challenges, we improve parameterisations of soft equivariance and learn the amount of equivariance in layers by optimising the marginal likelihood, estimated using differentiable Laplace approximations. The objective balances data fit and model complexity enabling layer-wise symmetry discovery in deep networks. We demonstrate the ability to automatically learn layer-wise equivariances on image classification tasks, achieving equivalent or improved performance over baselines with hard-coded symmetry. We demonstrate automatically learning layer-wise symmetry structure on image classification tasks. To do so, we improve upon existing parameterisations of differentiable equivariance and derive corresponding Kronecker-factored Laplace approximations to the marginal likelihood. On image classification, we show that our method automatically learns convolutional structure in early layers and achieves similar or improved performance compared to architectures with hard-coded symmetry.