Research map / Learning theory and algorithms / Optimal transport
SNEkhorn: Dimension Reduction with Symmetric Entropic Affinities
Hugues Van Assel, Titouan Vayer, Rémi Flamary, Nicolas Courty
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 symmetric entropic affinity P^se instead of projected entropic affinity P^e | spectral clustering · CuMiDa Throat | ARI x100: 11.8 -> 44.2 | Improved |
| use symmetric entropic affinity P^se instead of P^e | spectral clustering · SNAREseq | ARI x100: 55.4 -> 96.6 | Improved |
| use P^se instead of P^e | spectral clustering · CuMiDa Colorectal (2nd listed) | ARI x100: 94.0 -> 79.3 | Hurt |
| use P^se instead of P^e | spectral clustering · 12 genomics datasets | ARI: 8 out of 12 | Mixed |
| use P^se instead of self-tuning P^st | spectral clustering · scGEM | ARI x100: 74.8 -> 71.6 | Hurt |
| use t-SNEkhorn instead of t-SNE | DR on COIL · COIL | silhouette x100: 30.7 -> 52.3 | Improved |
| use t-SNEkhorn instead of t-SNE | DR on UMNIST · UMNIST | silhouette x100: -0.2 -> 25.4 | Improved |
| use t-SNEkhorn instead of t-SNE | DR on SNAREseq · SNAREseq | silhouette x100: 35.8 -> 67.9 | Improved |
| use t-SNEkhorn instead of t-SNE and UMAP | DR on 16 datasets · 16 datasets | silhouette: 15 of 16 datasets | Improved |
| use t-SNEkhorn instead of t-SNE and UMAP | DR on 16 datasets · 16 datasets | trustworthiness: 12 of 16 datasets | Mixed |
| use t-SNEkhorn instead of t-SNE | DR on CuMiDa Prostate · CuMiDa Prostate | trustworthiness x100: 80.6 -> 73.1 | Hurt |
| use symmetric entropic affinity instead of projected entropic affinity under heteroscedastic noise | affinity · synthetic multinomial toy data | cluster separation | Improved |
Closest papers by meaning
- On Probabilistic Embeddings in Optimal Dimension Reduction (JMLR 2025 · similarity 0.86)
- Structural-Entropy-Based Sample Selection for Efficient and Effective Learning (ICLR 2025 · similarity 0.85)
- A Geometry-Aware Efficient Algorithm for Compositional Entropic Risk Minimization (ICML 2026 · similarity 0.85)
- Simultaneous Clustering and Model Selection for Tensor Affinities (CVPR 2016 · similarity 0.85)
- On the Efficiency of Entropic Regularized Algorithms for Optimal Transport (JMLR 2022 · similarity 0.85)
- Accelerating Sinkhorn algorithm with sparse Newton iterations (ICLR 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.
