Research map / Probabilistic and scientific modeling / Variational and generative sampling
Practical and Consistent Estimation of f-Divergences
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 |
|---|---|---|---|
| Approximate the intractable aggregate distribution with a random mixture of N conditional distributions | f-divergence estimation with RAM · f-divergence estimation | Bias: KL: N^{-1}; TV: N^{-\frac{1}{2}}; χ²: N^{-1}; H²: N^{-\frac{1}{2}}; JS: N^{-\frac{1}{4}} | Helped |
| Increase the number N of sampled data points in the random mixture | RAM-MC · Synthetic normal distributions, d∈{1, 4, 16} | Bias and variance | Helped |
| Use RAM-MC instead of density-estimation baselines | Synthetic divergence estimation · Synthetic normal distributions, especially d=16 | Estimation accuracy | Helped |
| Increase Monte Carlo sample count M | RAM-MC · CelebA pretrained autoencoders | Variance | Helped |
| Estimate the divergence using the finite random mixture rather than estimating densities | RAM-MC · CelebA pretrained autoencoders | Estimation accuracy | Helped |
| Use a larger latent dimension | RAM-MC · Synthetic normal distributions | Bias | Hurt |
| Use an arbitrary importance-sampling proposal instead of a specifically optimized proposal | RAM-MC · f-divergence estimation | Estimator expectation | No clear effect |
| Use the Minibatch Weighted Sample entropy estimate with its log D offset | MWS · Entropy estimation | Consistency: log D | Hurt |
Closest papers by meaning
- Mutual Information Estimation via $f$-Divergence and Data Derangements (NeurIPS 2024 · similarity 0.88)
- Bounds on $L_p$ Errors in Density Ratio Estimation via $f$-Divergence Loss Functions (ICLR 2025 · similarity 0.87)
- Statistical and Geometrical properties of the Kernel Kullback-Leibler divergence (NeurIPS 2024 · similarity 0.85)
- $f$-Divergence Based Classification: Beyond the Use of Cross-Entropy (ICML 2024 · similarity 0.85)
- Robust Diffusion Models via Divergence-Induced Weighted Denoising (NeurIPS 2026 · similarity 0.85)
- Connecting Jensen–Shannon and Kullback–Leibler Divergences: A New Bound for Representation Learning (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.