Research map / Learning theory and algorithms / Optimal transport
Flattening the Density Gradient for Eliminating Spatial Centrality to Reduce Hubness
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 |
|---|---|---|---|
| subtract sample-wise centrality (squared distance to global centroid) from squared Euclidean distance (DisSimGlobal) | Euclidean distance k-NN · i.i.d. Gaussian(0,I), n=1000, d=1000 | N10 skewness: 2.83 -> 0.38 | Helped |
| use global centroid correction (DisSimGlobal) on non-isotropic or mixture Gaussian data (ablation) | Euclidean distance k-NN · non-isotropic Gaussian and Gaussian mixture, d=1000 | N10 skewness | No clear effect |
| use local centroid (kappa-NN mean) correction (DisSimLocal) instead of global centroid (ablation) | DisSimGlobal · non-isotropic Gaussian and Gaussian mixture, d=1000 | N10 skewness | Helped |
| apply DisSimLocal to Euclidean distance for k-NN classification | Euclidean distance k-NN · Leukemia, Lung Cancer, MFeat, ISOLET | k-NN accuracy | Helped |
Closest papers by meaning
- On Finding Hubs in High Dimensions with Sampling (AAAI 2025 · similarity 0.87)
- Quantifying Centrality for Complex Data (NeurIPS 2026 · similarity 0.83)
- On Convergence of Nearest Neighbor Classifiers over Feature Transformations (NeurIPS 2020 · similarity 0.82)
- Towards Fairer Centroids in K-means Clustering (AAAI 2024 · similarity 0.82)
- Efficient Constrained K-center Clustering with Background Knowledge (AAAI 2024 · similarity 0.81)
- Average Sensitivity of Hierarchical $k$-Median Clustering (ICML 2025 · similarity 0.81)
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.