Research map / Optimization / Nonconvex and smooth optimization
Sinkhorn Barycenter via Functional Gradient Descent
Zebang Shen, Zhenfu Wang, Alejandro Ribeiro, Hamed Hassani
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 Sinkhorn Descent instead of functional Frank-Wolfe | Sinkhorn barycenter of concentric ellipses · Concentric ellipses | convergence rate | Improved |
| Use Sinkhorn Descent with support size 20 instead of functional Frank-Wolfe with support size 600 | Sinkhorn barycenter of concentric ellipses · Concentric ellipses | solution quality | Improved |
| Use Sinkhorn Descent instead of functional Frank-Wolfe | Distribution sketching · Cheetah image | solution quality | Improved |
| Use Sinkhorn Descent instead of functional Frank-Wolfe | Distribution sketching · Cheetah image | convergence rate | Improved |
| Use Sinkhorn Descent for the high-dimensional barycenter problem | Barycenter of Gaussians · 5 isotropic Gaussians in R^100 | convergence iterations: less than 20 iterations | Improved |
| Use Sinkhorn Descent instead of functional Frank-Wolfe | Visual barycenter and sketching results · Ellipses and cheetah sketching | visual quality | Improved |
Closest papers by meaning
- Sinkhorn Barycenters with Free Support via Frank-Wolfe Algorithm (NeurIPS 2019 · similarity 0.91)
- Debiased Sinkhorn barycenters (ICML 2020 · similarity 0.87)
- Accelerating Sinkhorn algorithm with sparse Newton iterations (ICLR 2024 · similarity 0.87)
- Differential Properties of Sinkhorn Approximation for Learning with Wasserstein Distance (NeurIPS 2018 · similarity 0.87)
- Distributional Reinforcement Learning with Regularized Wasserstein Loss (NeurIPS 2024 · similarity 0.84)
- Mirror and Preconditioned Gradient Descent in Wasserstein Space (NeurIPS 2024 · similarity 0.84)
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.
