Light Entropic Optimal Transport on Riemannian Manifolds

arXiv:2610.03085v1 Announce Type: new Abstract: Entropic Optimal Transport (EOT) has become a practical framework for learning stochastic couplings between complex distributions, with applications in generative modeling and domain adaptation. However, most EOT solvers are designed for Euclidean spaces, while manifold extensions remain limited and often rely on costly iterative methods, simulated dynamics, or generic neural models that do not fully exploit the underlying geometry. We introduce ManifoldLightOT, a light approach for learning kernel-induced EOT couplings directly on common manifolds. Using the kernel form of the EOT solution, we construct geometry-specific Gibbs kernels together with compatible potential parameterizations for spheres, tori, SO(3), and SE(3). These choices yield closed-form normalization and directly sampleable conditional distributions. Our formulation naturally extends to products of manifolds, making it applicable to more complex geometries. The parameters of the potentials are optimized directly from samples using Monte Carlo estimates of the learning objective. Through synthetic and real-world experiments, we show that ManifoldLightOT often outperforms existing manifold OT methods while retaining direct sampling.

Related

• A Biconvex Formulation for Stable Transport of Mixture Models with a Unique Solution
• Optimal Transport for Network Comparison: A Review with Machine Learning Applications
• Reliable Modeling of Distribution Shifts via Displacement-Reshaped Optimal Transport
• EMS Coreset: An Efficient Expectation-Maximization Algorithm for Sinkhorn Coreset
• Learning Geometry and Topology via Multi-Chart Flows

Source: arXiv cs.LG | 2026-10-05