Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Discriminative Calibration: Check Bayesian Computation from Simulations and Flexible Classifier
Yuling Yao, Justin Domke
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 a learned classifier instead of Bonferroni-corrected SBC rank tests | Bayesian computation calibration on Gaussian posterior bias and variance corruptions · Gaussian posterior | test power | Improved |
| Add log approximate density as a statistically meaningful classifier feature (ablation) | Weighted binary discriminative calibration · Galaxy clustering, three cosmology models | classifier prediction quality | Improved |
| Add log approximate density as a statistically meaningful classifier feature (ablation) | Weighted binary discriminative calibration · Galaxy clustering, three cosmology models | divergence estimate tightness | Improved |
| Use an ensemble of five flows instead of a single flow (ablation) | Simulation-based inference for galaxy clustering · Galaxy clustering, three cosmology models | estimated divergence | Improved |
| Increase the number of MCMC iterations | NUTS posterior samples · SBI benchmark: SLCP, Gaussian linear, and Gaussian mixture | estimated JS divergence | Improved |
| Increase the number of posterior draws used for multiclass classification | Multiclass discriminative calibration · Gaussian simulations with biased posterior mean | estimated divergence versus KL limit | Improved |
Closest papers by meaning
- Testing Calibration in Nearly-Linear Time (NeurIPS 2024 · similarity 0.87)
- Reassessing How to Compare and Improve the Calibration of Machine Learning Models (ICLR 2025 · similarity 0.85)
- Calibration without labels in multiple testing (NeurIPS 2026 · similarity 0.84)
- Combining Priors with Experience: Confidence Calibration Based on Binomial Process Modeling (AAAI 2025 · similarity 0.84)
- Bayesian Adaptive Calibration and Optimal Design (NeurIPS 2024 · similarity 0.84)
- Truthful Calibration Errors for Multi-Class Prediction (NeurIPS 2026 · 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.
