Riemannian Networks over Full-Rank Correlation Matrices

Ziheng Chen, Xiaojun Wu, Bernhard Schölkopf, Nicu SebeICML 2026 · OpenReview

Abstract

Representations on the Symmetric Positive Definite (SPD) manifold have garnered significant attention across different applications. In contrast, the manifold of full-rank correlation matrices, a normalized alternative to SPD matrices, remains largely underexplored. This paper introduces Riemannian networks over the correlation manifold, leveraging five recently developed correlation geometries. We systematically extend basic layers, including Multinomial Logistic Regression (MLR), Fully Connected (FC), and convolutional layers, to these geometries. Besides, we present methods for accurate backpropagation for two correlation geometries. Experiments comparing our approach against existing SPD and Grassmannian networks demonstrate its effectiveness.
BibTeX
@inproceedings{chen2026riemannian,
  author = {Ziheng Chen and Xiaojun Wu and Bernhard Schölkopf and Nicu Sebe},
  title = {{Riemannian Networks over Full-Rank Correlation Matrices}},
  booktitle = {International Conference on Machine Learning},
  year = {2026},
  url = {https://openreview.net/forum?id=8k4om4zj5E},
}

Loading related content...

Figure