Research map / Learning theory and algorithms / Sample complexity
The Implicit Metropolis-Hastings Algorithm
Kirill Neklyudov, Evgenii Egorov, Dmitry P Vetrov
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 |
|---|---|---|---|
| Filter WPGAN samples with implicit Metropolis-Hastings using a Markov proposal | WPGAN · CIFAR-10 | Inception Score | Improved |
| Filter WPGAN samples with implicit Metropolis-Hastings using a Markov proposal | WPGAN · CIFAR-10 | FID | Improved |
| Learn the discriminator with conventional cross-entropy (CCE) | Independent-proposal implicit Metropolis-Hastings · CIFAR-10 and CelebA | Inception Score | Improved |
| Learn the discriminator with conventional cross-entropy (CCE) | Independent-proposal implicit Metropolis-Hastings · CIFAR-10 and CelebA | FID | Improved |
| Use Markov cross-entropy (MCE) instead of upper-bound loss (UB) (ablation) | Markov-proposal discriminator · Synthetic mixture of two Gaussians | TV distance | No clear effect |
| Use conventional cross-entropy (CCE) instead of UB or MCE (ablation) | Independent-proposal discriminator · Synthetic mixture of two Gaussians | TV distance | Improved |
| Use CCE rather than UB or MCE (ablation) | Independent-proposal discriminator · CIFAR-10 and CelebA | FID | Improved |
| Replace independent sampling with a Markov proposal at the same rejection rate (ablation) | WPGAN · CIFAR-10 | Inception Score | Improved |
| Replace independent sampling with a Markov proposal at the same rejection rate (ablation) | WPGAN · CIFAR-10 | FID | Improved |
Closest papers by meaning
- Metropolis-Hastings Generative Adversarial Networks (ICML 2019 · similarity 0.88)
- AC-Sampler: Accelerate and Correct Diffusion Sampling with Metropolis-Hastings Algorithm (ICLR 2026 · similarity 0.86)
- Denoising Diffusion Implicit Models (ICLR 2021 · similarity 0.85)
- Revisiting Unbiased Implicit Variational Inference (ICML 2025 · similarity 0.84)
- Continuous Semi-Implicit Models (ICML 2025 · similarity 0.83)
- Ai-sampler: Adversarial Learning of Markov kernels with involutive maps (ICML 2024 · similarity 0.83)
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.
