Research map / Learning theory and algorithms / Kernels and regression theory
MMD-Fuse: Learning and Combining Kernels for Two-Sample Testing Without Data Splitting
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 MMD-FUSE instead of MMDAgg | MMD two-sample testing · CIFAR-10 vs CIFAR-10.1 | power: 0.883 -> 0.937 | Helped |
| Use the normalised FUSE-N statistic rather than unnormalised FUSE-1 (ablation) | MMD-FUSE · Synthetic and real-world experiments | power | Helped |
| Increase the number of fused kernels (ablation) | MMD-FUSE · Perturbed Uniform, six one-dimensional perturbations with amplitude a=0.5 | power | No clear effect |
| Use median-bandwidth kernels instead of adaptive kernel selection | MMD-based two-sample tests · Synthetic experiments | power | Hurt |
| Use MMD-D instead of MMD-FUSE | MMD two-sample testing · Galaxy MNIST, increasing sample size | power | Helped |
| Use AutoML instead of MMD-FUSE as sample size increases | MMD two-sample testing · Synthetic experiments with increasing sample size | power | Hurt |
| Use SCF instead of spatial-domain ME | Permutation-based two-sample testing · Mixture of Gaussians | power | Mixed |
| Use the spatial-domain ME test instead of SCF | Permutation-based two-sample testing · Perturbed Uniform, d in {1,2} | power | Helped |
Closest papers by meaning
- Kernel Selection is Model Selection: A Unified Complexity-Penalised Approach for MMD Two-Sample Tests (NeurIPS 2026 · similarity 0.92)
- A permutation-free kernel two-sample test (NeurIPS 2022 · similarity 0.91)
- Kernel-based Maximum-of-difference Test for Two-sample Comparison (ICML 2026 · similarity 0.91)
- DUAL: Learning Diverse Kernels for Aggregated Two-sample and Independence Testing (NeurIPS 2025 · similarity 0.88)
- Optimal Algorithms for Augmented Testing of Discrete Distributions (NeurIPS 2024 · similarity 0.85)
- Kernel conditional tests from learning-theoretic bounds (NeurIPS 2025 · similarity 0.85)
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.