Provably Strict Generalisation Benefit for Invariance in Kernel Methods

Abstract

It is a commonly held belief that enforcing invariance improves generalisation. Although this approach enjoys widespread popularity, it is only very recently that a rigorous theoretical demonstration of this benefit has been established. In this work we build on the function space perspective of Elesedy and Zaidi arXiv:2102.10333 to derive a strictly non-zero generalisation benefit of incorporating invariance in kernel ridge regression when the target is invariant to the action of a compact group. We study invariance enforced by feature averaging and find that generalisation is governed by a notion of effective dimension that arises from the interplay between the kernel and the group. In building towards this result, we find that the action of the group induces an orthogonal decomposition of both the reproducing kernel Hilbert space and its kernel, which may be of interest in its own right.

arXiv commentAppeared in NeurIPS 2021

References in corpus (12)

• Neural Tangent Kernel: Convergence and Generalization in Neural Networks
• Group Equivariant Convolutional Networks
• On the Generalization of Equivariance and Convolution in Neural Networks to the Action of Compact Groups
• Equivariance Through Parameter-Sharing
• 3D G-CNNs for Pulmonary Nodule Detection
• Unsupervised Learning of Invariant Representations in Hierarchical Architectures
• Generalization Error of Invariant Classifiers
• Learning with invariances in random features and kernel models
• On the Benefits of Invariance in Neural Networks
• Learning with Group Invariant Features: A Kernel Perspective
• Improving Transformation Invariance in Contrastive Representation Learning
• Better, Faster Fermionic Neural Networks