Research map / Learning theory and algorithms / Optimal transport
Bringing regularized optimal transport to lightspeed: a splitting method adapted for GPUs
Jacob Lindbäck, Zesen Wang, Mikael Johansson
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 RDROT (Douglas-Rachford, GPU) instead of L-BFGS dual ascent for quadratic OT | QDROT · Quadratic-regularized OT, 1000x1000 and 2000x3000, 50 datasets | runtime to 1e-4 accuracy | Improved |
| Use GLDROT instead of Sinkhorn-based group-lasso OT (Ent 1e-3) | GLSK · Domain adaptation simulation | median runtime (s): 3.77 -> 0.0619 | Improved |
| Use GLDROT instead of GLSK (Ent 1e-3) for adaptation quality | GLSK · Domain adaptation simulation | median aggregated W2 distance: 0.311 -> 0.0576 | Improved |
| Increase entropic regularization in GLSK from 1e-3 to 1e-1 (ablation) | GLSK · Domain adaptation simulation | median aggregated W2 distance: 0.311 -> 8.24 | Hurt |
| Increase entropic regularization in GLSK from 1e-1 to 1e+2 (ablation) | GLSK · Domain adaptation simulation | median runtime (s): 3.73 -> 1.86 | Improved |
| Increase entropic regularization in GLSK from 1e-1 to 1e+2 (ablation) | GLSK · Domain adaptation simulation | median aggregated W2 distance: 8.24 -> 52.8 | Hurt |
| Increase group-lasso weight in GLDROT from 1e-3 to 5e-3 (ablation) | GLDROT · Domain adaptation simulation | median runtime (s): 0.0619 -> 0.0384 | Improved |
| Increase group-lasso weight in GLDROT from 1e-3 to 5e-3 (ablation) | GLDROT · Domain adaptation simulation | median aggregated W2 distance: 0.0576 -> 0.0879 | Hurt |
| Reduce Sinkhorn entropic regularization to limit shrinkage (ablation) | GLSK · Domain adaptation | runtime and stability | Mixed |
| Use GPU RDROT with low regularization for Minibatch Energy Distance GAN loss | RDROT · MNIST/CIFAR10 GAN training | training time | No clear effect |
Closest papers by meaning
- cuRegOT: A GPU-Accelerated Solver for Entropic-Regularized Optimal Transport (ICML 2026 · similarity 0.92)
- A Truncated Newton Method for Optimal Transport (ICLR 2025 · similarity 0.90)
- Log-Averaged Mirror Prox for Fast, Large-Scale Optimal Transport in Linear Space (NeurIPS 2026 · similarity 0.87)
- Decreasing Entropic Regularization Averaged Gradient for Semi-Discrete Optimal Transport (NeurIPS 2025 · similarity 0.85)
- Expectile Regularization for Fast and Accurate Training of Neural Optimal Transport (NeurIPS 2024 · similarity 0.84)
- On Partial Optimal Transport: Revising the Infeasibility of Sinkhorn and Efficient Gradient Methods (AAAI 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.
