Research map / Learning theory and algorithms / Optimal transport
ITSPACE: Monotone Gaussian Optimal Transport Updates
Woojoo Na, Jennifer Dy
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 ITSPACE factor update instead of BW gradient descent | rank-16 covariance alignment · Camelyon17 time to gap<=0.1 | time (s): 6.85 -> 0.0214 | Improved |
| use ITSPACE instead of Sinkhorn variants under BW evaluation | rank-16 covariance alignment · Camelyon17, VisDA-2017, Terra gap<=0.02 | reaches threshold within K=20: Sinkhorn variants NR; ITSPACE reaches it | Improved |
| use ITSPACE instead of Log-Euclidean or AIRM updates with projection | rank-16 covariance alignment · Camelyon17 time to gap<=0.1 | time (s): 16.4 -> 0.0214 (Log-Euclid.), 19 -> 0.0214 (AIRM) | Improved |
| adapt covariance with ITSPACE at K=1 versus no adaptation | frozen linear head · Camelyon17 CovDrift-MR | AUROC: 70.22 -> 73.94 | Improved |
| adapt covariance with ITSPACE K=5 versus no adaptation | frozen linear head · VisDA-2017 CovDrift-MR | accuracy: 75.68 -> 84.26 | Improved |
| adapt covariance with ITSPACE K=5 versus no adaptation | frozen linear head · Terra CovDrift-MR | accuracy: 58.89 -> 63.94 | Improved |
| use ITSPACE instead of BW-GD at K=1 | downstream drift correction · Camelyon17 CovDrift-MR | AUROC: 73.97 -> 73.94 | No clear effect |
| use ITSPACE instead of Sinkhorn at K=5 | downstream drift correction · Terra CovDrift-MR | accuracy: 65.00 -> 63.94 | Hurt |
| use ITSPACE instead of BW-geodesic at K=1 | downstream drift correction · Camelyon17 CovDrift-MR | AUROC: 71.92 -> 73.94 | Improved |
| increase update budget from K=2 to K=20 for ITSPACE (ablation) | ITSPACE · Camelyon17 CovDrift-MR | AUROC: 74.01 -> 74.01 | No clear effect |
| use ITSPACE instead of BW-geodesic regarding adaptation time at K=5 | downstream drift correction · Camelyon17 CovDrift-MR | t5 (s): 0.015 -> 0.035 | Hurt |
Closest papers by meaning
- Optimal Tensor Transport (AAAI 2022 · similarity 0.88)
- Geometry-Aware Optimal Transport: Fast Intrinsic Dimension and Wasserstein Distance Estimation (NeurIPS 2026 · similarity 0.86)
- Hierarchical Refinement: Optimal Transport to Infinity and Beyond (ICML 2025 · similarity 0.86)
- Low-Rank Optimal Transport through Factor Relaxation with Latent Coupling (NeurIPS 2024 · similarity 0.86)
- Joint Metric Space Embedding by Unbalanced Optimal Transport with Gromov–Wasserstein Marginal Penalization (ICML 2025 · similarity 0.86)
- Compress to Impress: Efficient LLM Adaptation Using a Single Gradient Step on 100 Samples (NeurIPS 2025 · 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.
