arXiv AI

Ollivier-Ricci Curvature of Riemannian Manifolds and Directed Graphs with Applications to Graph Neural Networks

arXiv Machine Learning
Jul 27

Local-Global Geometric Insights for Graph Neural Networks via Entropic Curvature

arXiv:2607. 22381v1 Announce Type: new Abstract: Curvature notions on graphs, particularly Ollivier-Ricci and Forman, have emerged as powerful tools for addressing fundamental issues in Graph Neural Networks (GNNs) such as oversmoothing and oversquashing, but rely almost exclusively on local edge-level comparisons and therefore fail to certify how information actually propagates over long distances.

By Rachid Caich, Yassine Abbahaddou
arXiv Machine Learning
6d ago

Optimal Transport for Network Comparison: A Review with Machine Learning Applications

The paper reviews the use of optimal transport for comparing undirected, unweighted graphs, focusing on three main distances: Wasserstein, Gromov-Wasserstein, and Bures-Wasserstein. It discusses closed-form solutions for the Wasserstein distance in one dimension, how transport plans identify influential nodes after perturbations, and derives spectral bounds for the Bures-Wasserstein distance to avoid full decompositions. The authors evaluate these distances on synthetic clustering data and a real-world time‑series network for anomaly detection.

By James Hyun, Fran\c{c}ois G. Meyer
arXiv Machine Learning
Aug 19

Network Denoising Revisited: A Ricci-Flow-Inspired Graph Diffusion Method

The paper introduces Ricci-Diffusion, a graph denoising technique that uses curvature-guided diffusion inspired by Ricci flow. Unlike traditional similarity-driven methods, it modulates local transport in the diffusion kernel based on edge-level curvature, steering edge-weight updates toward a more regular graph geometry. The authors provide theoretical analysis showing curvature’s ability to distinguish graph structures and induce first-order corrections, and demonstrate that the method converges to a stable denoised network, improving structure recovery and downstream performance on real-world and synthetic graphs.

By Ye Fang, Chuan-Xian Ren
arXiv AI
Jul 22

Riemannian Deep Learning:Modules, Networks, and Geometries

arXiv:2607. 19305v1 Announce Type: cross Abstract: Deep neural networks on manifold-valued representations have attracted growing interest, but many basic components remain tied to specific manifolds, rely on Euclidean approximations, or require costly and numerically fragile geometric operations.

By Chen Ziheng
arXiv AI
Jul 24

Riemannian Deep Learning: Modules, Networks, and Geometries

arXiv:2607. 19305v2 Announce Type: replace-cross Abstract: Deep neural networks on manifold-valued representations have attracted growing interest, but many basic components remain tied to specific manifolds, rely on Euclidean approximations, or require costly and numerically fragile geometric operations.

By Chen Ziheng