Riemannian batch normalization for SPD neural networks

TL;DR

This paper introduces a Riemannian batch normalization technique for SPD neural networks, leveraging geometric operations on the manifold to improve classification performance and robustness across diverse data types.

Contribution

It proposes a novel Riemannian batch normalization layer and a manifold-constrained gradient descent algorithm for SPD matrices, enhancing deep learning on structured data.

Findings

Improved classification accuracy across multiple datasets

Enhanced robustness to limited data scenarios

Effective integration of Riemannian geometry in deep learning

Abstract

Covariance matrices have attracted attention for machine learning applications due to their capacity to capture interesting structure in the data. The main challenge is that one needs to take into account the particular geometry of the Riemannian manifold of symmetric positive definite (SPD) matrices they belong to. In the context of deep networks, several architectures for these matrices have recently been proposed. In our article, we introduce a Riemannian batch normalization (batchnorm) algorithm, which generalizes the one used in Euclidean nets. This novel layer makes use of geometric operations on the manifold, notably the Riemannian barycenter, parallel transport and non-linear structured matrix transformations. We derive a new manifold-constrained gradient descent algorithm working in the space of SPD matrices, allowing to learn the batchnorm layer. We validate our proposed approach with experiments in three different contexts on diverse data types: a drone recognition dataset from radar observations, and on emotion and action recognition datasets from video and motion capture data. Experiments show that the Riemannian batchnorm systematically gives better classification performance compared with leading methods and a remarkable robustness to lack of data.