Research map / Optimization / Nonconvex and smooth optimization
Mirror Sinkhorn: Fast Online Optimization on Transport Polytopes
Marin Ballu, Quentin Berthet
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use a single Sinkhorn normalization step per gradient update (single-loop) instead of multiple nested Sinkhorn steps (k_S=10) (ablation) | Mirror descent approximation for strongly convex objective · Strongly convex synthetic (m=50, n=60) | convergence of f(gamma_t)-f(gamma*) | No clear effect |
| Increase Sinkhorn steps per gradient update from 1 to k_S=10 (nested loop) (ablation) | Mirror descent approximation · Strongly convex synthetic and Wasserstein Procrustes | convergence speed per algorithmic step | Hurt |
| Use stochastic gradients (additive noise) instead of exact gradients | Mirror Sinkhorn · Synthetic OT and strongly convex objective | convergence of objective | No clear effect |
| Use the Sinkhorn-style constant regularization alpha vs. Mirror Sinkhorn (no entropic bias) on OT | Sinkhorn baseline vs Mirror Sinkhorn on OT · Synthetic OT (m=n=100) | regularization bias / convergence of <C,gamma_t> | Improved |
| Larger entropic regularization alpha in Sinkhorn (further from boundary) | Sinkhorn baseline · Synthetic OT (m=n=100) | marginal constraint violation decay | Hurt |
Closest papers by meaning
- On Efficient Optimal Transport: An Analysis of Greedy and Accelerated Mirror Descent Algorithms (ICML 2019 · similarity 0.89)
- Accelerating Sinkhorn algorithm with sparse Newton iterations (ICLR 2024 · similarity 0.88)
- On Partial Optimal Transport: Revising the Infeasibility of Sinkhorn and Efficient Gradient Methods (AAAI 2024 · similarity 0.87)
- Safe and Sparse Newton Method for Entropic-Regularized Optimal Transport (NeurIPS 2024 · similarity 0.86)
- Log-Averaged Mirror Prox for Fast, Large-Scale Optimal Transport in Linear Space (NeurIPS 2026 · similarity 0.86)
- Geometry-Aware Optimal Transport: Fast Intrinsic Dimension and Wasserstein Distance Estimation (NeurIPS 2026 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
