Research map / Probabilistic and scientific modeling / Variational and generative sampling
Fast ε-free Inference of Simulation Models with Bayesian Conditional Density Estimation
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 |
|---|---|---|---|
| Train the MDN with a learned proposal prior instead of the actual prior (ablation) | MDN posterior estimator · Mixture of two Gaussians | number of simulations: 10K -> 1000 | Helped |
| Sequentially fit a proposal prior rather than training with prior samples (ablation) | MDN · Bayesian linear regression | simulation cost and KL divergence | Helped |
| Use MDN methods instead of Monte Carlo ABC | posterior inference · Lotka-Volterra | neg. log probability of true parameters | Helped |
| Use more than one Gaussian component in the MDN (ablation) | MDN · Lotka-Volterra | posterior quality | No clear effect |
| Train MDN with proposal prior instead of prior (smaller network suffices) (ablation) | MDN · Lotka-Volterra | network size and simulations: 2 hidden layers -> 1 hidden layer of 50 tanh units | Helped |
| Train MDN with proposal prior; less capacity needed (ablation) | MDN · M/G/1 queue | final posterior accuracy and compute | Helped |
| Use MDN-SVI instead of conventional MDN for proposal prior learning | Algorithm 1 · general | robustness and simulation efficiency | Helped |
| Use MDN methods instead of ABC | posterior inference · M/G/1 queue | neg. log probability of true parameters | Helped |
| Use MDN methods instead of Monte Carlo ABC | posterior inference · Bayesian linear regression | simulations vs KL divergence | Helped |
| Very long run of SMC-ABC vs MDN with proposal | SMC-ABC · Bayesian linear regression | KL divergence | No clear effect |
Closest papers by meaning
- Learning Likelihood-Free Reference Priors (ICML 2025 · similarity 0.91)
- Approximate Bayesian Computation with Domain Expert in the Loop (ICML 2022 · similarity 0.91)
- DR-ABC: Approximate Bayesian Computation with Kernel-Based Distribution Regression (ICML 2016 · similarity 0.88)
- Automated Model Selection with Bayesian Quadrature (ICML 2019 · similarity 0.88)
- Consistency Models for Scalable and Fast Simulation-Based Inference (NeurIPS 2024 · similarity 0.87)
- Effortless, Simulation-Efficient Bayesian Inference using Tabular Foundation Models (NeurIPS 2025 · similarity 0.87)
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.