Research map / Probabilistic and scientific modeling / Variational and generative sampling
Reliable Estimation of KL Divergence using a Discriminator in Reproducing Kernel Hilbert Space
Sandesh Ghimire, Aria Masoomi, Jennifer Dy
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 an RKHS discriminator with complexity control instead of an unconstrained neural network discriminator | KL estimation · 2D Gaussian KL (true KL 1.3) | estimation error: 0.5 to 0.04 | Improved |
| Use an RKHS discriminator with complexity control instead of an unconstrained neural network discriminator | KL estimation · 2D Gaussian KL (true KL 1.3) | variance: 0.2 to 0.002 | Improved |
| Use an RKHS discriminator with complexity control instead of an unconstrained neural network discriminator | KL estimation · 2D Gaussian KL (true KL 13.8) | estimation error: 5.8 to 1.07 | Improved |
| Use an RKHS discriminator with complexity control instead of an unconstrained neural network discriminator | KL estimation · 2D Gaussian KL (true KL 13.8) | variance: 223 to 4.4 | Improved |
| Use an RKHS discriminator with complexity control instead of an unconstrained neural network discriminator | KL estimation · 2D Gaussian KL (true KL 38.29) | estimation error: 60.6 to 9.7 | Improved |
| Use an RKHS discriminator with complexity control instead of an unconstrained neural network discriminator | KL estimation · 2D Gaussian KL (true KL 38.29) | variance: 3521 to 33 | Improved |
| Increase RKHS norm regularization (ablation) | RKHS discriminator · 2D Gaussian KL estimation | variance | Improved |
| Increase RKHS norm regularization (ablation) | RKHS discriminator · 2D Gaussian KL estimation | bias | Mixed |
| Use an RKHS discriminator with complexity control instead of the basic neural-network discriminator | Adversarial Variational Bayes · MNIST | training stability | Improved |
| Use an RKHS discriminator with complexity control instead of an unconstrained neural network discriminator | Mutual information estimation · 20-d Gaussian distributions; mutual information from 2 to 10 | mutual information estimation performance | Mixed |
| Increase RKHS norm regularization (ablation) | RKHS discriminator · Mutual information estimation on 20-d Gaussian distributions | variance | Improved |
| Increase RKHS norm regularization (ablation) | RKHS discriminator · Mutual information estimation on 20-d Gaussian distributions | bias | Hurt |
Closest papers by meaning
- Statistical and Geometrical properties of the Kernel Kullback-Leibler divergence (NeurIPS 2024 · similarity 0.92)
- Decoupled Kullback-Leibler Divergence Loss (NeurIPS 2024 · similarity 0.87)
- Connecting Jensen–Shannon and Kullback–Leibler Divergences: A New Bound for Representation Learning (NeurIPS 2025 · similarity 0.86)
- Better Estimation of the Kullback--Leibler Divergence Between Language Models (NeurIPS 2025 · similarity 0.84)
- Mutual Information Estimation via $f$-Divergence and Data Derangements (NeurIPS 2024 · similarity 0.84)
- Cauchy-Schwarz Divergence Information Bottleneck for Regression (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.
