Research map / Learning theory and algorithms / Optimal transport
A Memory-Efficient Hierarchical Algorithm for Large-scale Optimal Transport Problems
Wenzhou Xia, Ya-Nan Zhu, Jingwei Liang, Xiaoqun Zhang
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 |
|---|---|---|---|
| add GPU PDHG solver cuPDLPx in place of Gurobi barrier (ablation) | HALO · DOTmark r=256 | time (s): 159.06 -> 4.31 | Improved |
| add multiscale hierarchy (ablation) | HALO · DOTmark r=64 | time (s): 128.4 -> 1.50 | Improved |
| add dual-violation correction to updateActive (ablation) | HALO · DOTmark r=1024 | max time (s): 633.34 -> 154.20 | Improved |
| add dual-violation correction to updateActive (ablation) | HALO · DOTmark r=1024 | avg time (s): 52.85 -> 27.73 | Improved |
| use constant stepsize via Pock-Chambolle rescaling instead of power iteration (ablation) | HALO with cuPDLPx · DOTmark r=1024 | speedup: 1.97× | Improved |
| raise shielding candidate budget Umax from 10 to 20 (ablation) | HALO · ModelNet10 n=2^18 | time (s): 265.8 -> 229.7 | Improved |
| HALO versus M3S entropic solver at r=1024 | HALO · DOTmark r=1024 | time (s): 247.22 -> 27.73 | Improved |
| HALO versus HOT at r=512 | HALO · DOTmark r=512 | time (s): 78.43 -> 11.17 | Improved |
| HALO versus HiRef at n=2^18 | HALO · ModelNet10 n=2^18 | time (s): 422.7 -> 229.7 | Improved |
| HALO versus ShortCut at r=512 | HALO · DOTmark r=512 | time (s): 438.14 -> 11.17 | Improved |
| HALO versus ShortCut in memory at r=512 | HALO · DOTmark r=512 | peak memory (GB): 0.56 -> 2.07 | Hurt |
| sparser image classes (Shapes, Microscopy) versus dense ClassicImages | HALO · DOTmark r=1024 | time (s): 23.78 -> 44.39 | Hurt |
Closest papers by meaning
- A Truncated Newton Method for Optimal Transport (ICLR 2025 · similarity 0.88)
- On Partial Optimal Transport: Revising the Infeasibility of Sinkhorn and Efficient Gradient Methods (AAAI 2024 · similarity 0.87)
- Log-Averaged Mirror Prox for Fast, Large-Scale Optimal Transport in Linear Space (NeurIPS 2026 · similarity 0.87)
- Hierarchical Refinement: Optimal Transport to Infinity and Beyond (ICML 2025 · similarity 0.87)
- Geometry-Aware Optimal Transport: Fast Intrinsic Dimension and Wasserstein Distance Estimation (NeurIPS 2026 · similarity 0.85)
- A Provably Convergent and Practical Algorithm for Gromov–Wasserstein Optimal Transport (NeurIPS 2026 · similarity 0.85)
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.
