Research map / Learning theory and algorithms / Optimal transport
Generalizing Point Embeddings using the Wasserstein Space of Elliptical Distributions
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 |
|---|---|---|---|
| Use elliptical Wasserstein-Bures embeddings instead of diagonal Gaussian embeddings | Word similarity embeddings · WordSim, WordSim-R, WordSim-S, MEN, MC, SimLex, RG, YP, MT-287, MT-771, RW | Spearman rank correlation: WordSim 53.45 -> 66.02; SimLex 25.09 -> 24.09; RG 69.77 -> 65.58 | Mixed |
| Use elliptical Wasserstein-Bures embeddings instead of diagonal Gaussian cosine embeddings | Entailment embeddings · Entailment | F1: 0.74 -> 0.73 | Hurt |
| Use elliptical Wasserstein-Bures embeddings instead of diagonal Gaussian embeddings | Entailment embeddings · Entailment | average precision: 0.70 -> 0.70 | No clear effect |
| Use elliptical embeddings for MDS rather than point-based standard MDS | MDS visualization · PISA scores, 35 OECD countries | normalized stress: 0.62 -> 5e-3 | Helped |
| Use separately normalized mean and covariance cosine terms instead of jointly normalized Wasserstein-Bures cosine (ablation) | Elliptical word similarity embeddings · Word similarity benchmarks | Pearson rank correlation | Helped |
Closest papers by meaning
- Fourier Sliced-Wasserstein Embedding for Multisets and Measures (ICLR 2025 · similarity 0.89)
- Aligning Embeddings and Geometric Random Graphs: Informational Results and Computational Approaches for the Procrustes-Wasserstein Problem (NeurIPS 2024 · similarity 0.86)
- A New Robust Partial p-Wasserstein-Based Metric for Comparing Distributions (ICML 2024 · similarity 0.86)
- Weighted Embeddings for Low-Dimensional Graph Representation (AAAI 2025 · similarity 0.86)
- Kernel Quantile Embeddings and Associated Probability Metrics (ICML 2025 · similarity 0.86)
- Hyperbolic Embeddings of Supervised Models (NeurIPS 2024 · similarity 0.86)
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.