Research map / Learning theory and algorithms / Kernels and regression theory
Learning to Approximate a Bregman Divergence
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 max-affine Bregman divergence learning instead of Mahalanobis metric learning | synthetic divergence regression · KL, LogDet, and Itakura-Saito divergence regression | fit to target divergence | Helped |
| Use max-affine Bregman divergence learning instead of Mahalanobis metric learning | PBDL · UCI clustering and ranking benchmarks | benchmark performance | Mixed |
| Use K<n with farthest-point clustering rather than K=n (ablation) | PBDL · UCI metric-learning experiments | benchmark performance | No clear effect |
| Use max-affine generating functions rather than a nonnegative sum of fixed convex basis functions | Bregman divergence learning · benchmark metric learning | accuracy | Helped |
Closest papers by meaning
- Deep Divergence Learning (ICML 2020 · similarity 0.93)
- Neural Bregman Divergences for Distance Learning (ICLR 2023 · similarity 0.92)
- A scaled Bregman theorem with applications (NeurIPS 2016 · similarity 0.91)
- Difference-of-submodular Bregman Divergence (ICLR 2025 · similarity 0.89)
- Learning Bregman Divergences with Application to Robustness (NeurIPS 2024 · similarity 0.85)
- Relating Misfit to Gain in Weak-to-Strong Generalization Beyond the Squared Loss (ICML 2025 · 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.