aiqbench analysis
CP4SBI addresses a critical limitation in simulation-based inference by providing a model-agnostic framework that corrects miscalibration in neural posterior estimators. By leveraging regression trees or CDF-based methods, it ensures finite-sample local coverage guarantees for various scoring functions, directly improving uncertainty quantification in complex non-linear models. This advancement is particularly relevant for researchers relying on score-diffusion modeling and normalizing flows where accurate credible sets are essential.
CP4SBI: Local Conformal Calibration of Credible Sets in Simulation-Based Inference
arXiv:2508.17077v3 Announce Type: replace-cross Abstract: Current experimental scientists have been increasingly relying on simulation-based inference (SBI) to invert complex non-linear models with intractable likelihoods. However, posterior approximations obtained with SBI are often miscalibrated, causing credible regions to undercover true parameters. We develop $\texttt{CP4SBI}$, a model-agnostic conformal calibration framework that constructs credible sets with local Bayesian coverage. Our two proposed variants, namely local calibration via regression trees and CDF-based calibration, enable finite-sample local coverage guarantees for any scoring function, including HPD, symmetric, and quantile-based regions. Experiments on widely used SBI benchmarks demonstrate that our approach improves the quality of uncertainty quantification for neural posterior estimators using both normalizing flows and score-diffusion modeling.
Abstract:Current experimental scientists have been increasingly relying on simulation-based inference (SBI) to invert complex non-linear models with intractable likelihoods. However, posterior approximations obtained with SBI are often miscalibrated, causing credible regions to undercover true parameters. We develop $\texttt{CP4SBI}$, a model-agnostic conformal calibration framework that constructs credible sets with local Bayesian coverage. Our two proposed variants, namely local calibration via regression trees and CDF-based calibration, enable finite-sample local coverage guarantees for any scoring function, including HPD, symmetric, and quantile-based regions. Experiments on widely used SBI benchmarks demonstrate that our approach improves the quality of uncertainty quantification for neural posterior estimators using both normalizing flows and score-diffusion modeling.
Machine Learning (stat.ML); Machine Learning (cs.LG)
arXiv:2508.17077 [stat.ML]
(or arXiv:2508.17077v3 [stat.ML] for this version)
Full text under the publisher's open licence · 120 words
