Research map / Probabilistic and scientific modeling / Variational and generative sampling
Intrinsic Sliced Wasserstein Distances for Comparing Collections of Probability Distributions on Manifolds and Graphs
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Intrinsic slicing (ISW2) instead of extrinsic Euclidean slicing (SW) | two-set distribution comparison on a graph · 200-gon graph synthetic | power: 1.00 vs 0.128 | Helped |
| Using MSW (max-sliced) instead of averaged slicing | two-set distribution comparison · 200-gon graph synthetic | size: size 0.63 vs nominal 0.05 | Hurt |
| Sobolev transport and MSW use permutation testing vs ISW2 embedding tests | runtime · 200-gon graph synthetic | computation time: ST 21474.88 s, MSW 61764.22 s vs ISW2 630.67 s | Hurt |
| Harmonic p-value combination of coordinate-wise t-tests (ISD comb) vs bootstrap T statistic (ablation) | finite interval synthetic · finite interval synthetic | power | Helped |
| Unsliced 24-dim inverse CDF embedding vs sliced ISW2 with only first eigenfunction (ablation) | finite interval synthetic · finite interval synthetic, D=24 | power | No clear effect |
| Increasing number of eigenfunctions L with fixed D=24 (slicing over more eigenfunctions) (ablation) | finite interval synthetic · finite interval synthetic | power | Mixed |
| Permutation null vs bootstrap for ISW2 T statistic | ISW2 testing · general | validity | No clear effect |
| Cauchy combination instead of harmonic mean p-value combination (ablation) | ISW2 p-value combination · general | validity | Hurt |
| Higher number of quantile knots D' (fixed L) in cylindrical data (ablation) | ISW2 embedding on cylinder · cylindrical synthetic | power | Helped |
| Choice of L on cylindrical data (ablation) | ISW2 embedding · cylindrical synthetic | power | No clear effect |
Closest papers by meaning
- Revisiting Tree-Sliced Wasserstein Distance Through the Lens of the Fermat–Weber Problem (ICLR 2026 · similarity 0.91)
- Slicing Wasserstein over Wasserstein via Functional Optimal Transport (ICLR 2026 · similarity 0.89)
- Fast Wasserstein rates for estimating probability distributions of probabilistic graphical models (NeurIPS 2026 · similarity 0.88)
- Tree-Sliced Wasserstein Distance with Nonlinear Projection (ICML 2025 · similarity 0.87)
- Statistical and Computational Guarantees of Kernel Max-Sliced Wasserstein Distances (ICML 2025 · similarity 0.87)
- A New Robust Partial p-Wasserstein-Based Metric for Comparing Distributions (ICML 2024 · similarity 0.87)
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.