Research map / Probabilistic and scientific modeling / Variational and generative sampling
Learning Manifold Implicitly via Explicit Heat-Kernel Learning
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 |
|---|---|---|---|
| learn kernel by heat-kernel framework in SVGD (HK-SVGD vs SVGD) | SVGD · UCI Concrete | test RMSE: 5.027 -> 4.814 | Helped |
| learn kernel by heat-kernel framework in SVGD | SVGD · UCI Wine | test RMSE: 0.645 -> 0.624 | Helped |
| learn kernel by heat-kernel framework in SVGD | SVGD · UCI Combined | test RMSE: 4.088 -> 4.077 | No clear effect |
| apply heat-kernel learning to matrix-valued SVGD (MSVGD-m) | MSVGD-m · UCI Protein | test RMSE: 3.852 -> 3.762 | Helped |
| apply heat-kernel learning to matrix-valued SVGD (MSVGD-m) | MSVGD-m · UCI Wine | test RMSE: 0.637 -> 0.629 | Helped |
| apply heat-kernel learning to matrix-valued SVGD (MSVGD-m) | MSVGD-m · UCI Wine | test log-likelihood: -0.988 -> -1.013 | Hurt |
| replace particles with parameter generator (HK-ISVGD) (ablation) | HK-SVGD · UCI Year | test RMSE: 8.656 -> 8.611 | Helped |
| replace particles with parameter generator (HK-ISVGD) (ablation) | HK-SVGD · UCI Concrete | test RMSE: 4.814 -> 4.824 | No clear effect |
| use learned heat kernel in MMD-GAN (HK) vs repulsive | MMD-GAN · CelebA | FID: 10.5 -> 9.7 | Helped |
| add kernel DK variant (HK-DK) over HK | MMD-GAN · CIFAR-10 DC-GAN | FID: 14.9 -> 13.2 | Helped |
| add kernel DK variant (HK-DK) over HK | MMD-GAN · CIFAR-10 ResNet | FID: 11.5 -> 10.3 | Helped |
| add kernel DK variant (HK-DK) over HK | MMD-GAN · ImageNet | FID: 29.6 -> 29.2 | Helped |
Closest papers by meaning
- Learning Distributions on Manifolds with Free-Form Flows (NeurIPS 2024 · similarity 0.87)
- Residual Deep Gaussian Processes on Manifolds (ICLR 2025 · similarity 0.86)
- Generative Modeling on Manifolds Through Mixture of Riemannian Diffusion Processes (ICML 2024 · similarity 0.86)
- Manifold Random Features (NeurIPS 2026 · similarity 0.85)
- Kernel-Gradient Drifting Models (NeurIPS 2026 · similarity 0.84)
- Curvature Enhanced Data Augmentation for Regression (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.