Research map / Probabilistic and scientific modeling / Variational and generative sampling
Likelihood-free MCMC with Amortized Approximate Ratio Estimators
Joeri Hermans, Volodimir Begy, Gilles Louppe
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 |
|---|---|---|---|
| replace a fixed reference hypothesis with dependent-versus-independent sample-parameter pairs to estimate the likelihood-to-evidence ratio (ablation) | AALR-MCMC · Tractable problem | MMD: 0.53 ± 0.004 -> 0.05 ± 0.005 | Improved |
| replace a fixed reference hypothesis with dependent-versus-independent sample-parameter pairs to estimate the likelihood-to-evidence ratio (ablation) | AALR-MCMC · Tractable problem | ROC AUC: 0.99 ± 0.0001 -> 0.58 ± 0.0080 | Improved |
| use the log-ratio output before the sigmoid rather than extracting the ratio from the sigmoid output (ablation) | AALR-MCMC ratio estimator · Circle simulation capacity experiment | gradient saturation | Improved |
| increase weight decay regularization (ablation) | AALR-MCMC ratio estimator · Tractable problem | posterior approximation variance | Hurt |
| increase ratio-estimator capacity from a fully connected network to LeNet and ResNet-18 (ablation) | AALR-MCMC ratio estimator · Circle simulation capacity experiment | posterior sharpness | Improved |
| use a low-capacity ratio estimator (ablation) | AALR-MCMC ratio estimator · Population model sequential ratio estimation | ROC AUC: .99, .92, .54, and finally .50 | Mixed |
| use AALR-MCMC instead of SNL | posterior inference · Tractable problem | MMD: 0.11 ± 0.091 -> 0.05 ± 0.005 | Improved |
Closest papers by meaning
- Direct Amortized Likelihood Ratio Estimation (AAAI 2024 · similarity 0.88)
- Leveraging Self-Consistency for Data-Efficient Amortized Bayesian Inference (ICML 2024 · similarity 0.87)
- Surrogate Likelihoods for Variational Annealed Importance Sampling (ICML 2022 · similarity 0.87)
- Approximate Bayesian inference from noisy likelihoods with Gaussian process emulated MCMC (JMLR 2024 · similarity 0.87)
- Fast ε-free Inference of Simulation Models with Bayesian Conditional Density Estimation (NeurIPS 2016 · similarity 0.87)
- Learning Likelihood-Free Reference Priors (ICML 2025 · similarity 0.86)
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.
