研究人员开发了一个新的黎曼几何框架，以改进低秩最优输运（OT）求解器。该方法将因子耦合建模为子流形，并使用 Fisher-Rao 乘积度量来推导出有效的投影器和收缩。该框架扩展到各种 OT 问题，包括线性 OT 和 Gromov-Wasserstein，提供每迭代线性的复杂度和全局最优性的证书。实验表明，与现有方法相比，收敛速度更快，性能更优。 AI
影响 引入了一种新颖的几何方法来优化机器学习算法，可能为复杂数据问题带来更有效和准确的解决方案。
排序理由 该集群包含两篇相同的 arXiv 研究论文，详细介绍了用于优化机器学习算法的新数学框架。
AI 生成摘要 · Google Gemini · 来自 4 个来源。 我们如何撰写摘要 →
arXiv:2606.12120v1 Announce Type: new Abstract: Low-rank optimal transport (OT) mitigates the quadratic scaling of classical solvers, yet existing approaches rely heavily on first-order mirror-descent updates that require careful hyperparameter tuning and ignore the optimization …
Low-rank optimal transport (OT) mitigates the quadratic scaling of classical solvers, yet existing approaches rely heavily on first-order mirror-descent updates that require careful hyperparameter tuning and ignore the optimization landscape's curvature. To address these limitati…
arXiv:2606.07926v1 Announce Type: new Abstract: Optimal transport couplings are probabilistic objects, while many learning pipelines require deterministic maps. In Euclidean space, barycentric projection converts a coupling into a map by taking conditional expectations, but on a …
Optimal transport couplings are probabilistic objects, while many learning pipelines require deterministic maps. In Euclidean space, barycentric projection converts a coupling into a map by taking conditional expectations, but on a Riemannian manifold curvature and cut loci make …