analiza aiqbench
This paper uses a conditional flow-matching model to learn amortized posteriors for simulator parameters, then applies conformal risk control to make the resulting prediction bands cover the target observations at a set rate. The reported gains are a synthetic benchmark with bands about 20% the width of KOH-MCMC bands, plus a measles SEIR case study where the failures expose structural model inadequacy. For ML readers, the useful point is the finite-sample coverage guarantee under exchangeability, which is a property that amortized inference methods usually lack, though the results still rest on synthetic setups and that assumption.
Prediction Uncertainty Controlled Computer Model Calibration
arXiv:2610.31163v1 Announce Type: cross Abstract: Computer-model calibration is usually undertaken in service of prediction, yet neither Bayesian Kennedy-O'Hagan (KOH) calibration nor amortized simulation-based inference is built to control the predictive error of the calibrated simulator. When parameters are weakly identified or the data-model discrepancy is misspecified, the simulator-output bands induced by a nominal posterior level can be far from nominal. We propose a two-layer framework that targets predictive miscoverage directly. First, a conditional flow-matching model learns an amortized posterior for the calibration parameters, trained on simulator outputs perturbed with Gaussian-process discrepancies of varying structure. Second, conformal risk control arranges posterior draws for the target observation into nested candidate sets and selects the smallest one whose fixed-rank pointwise bands attain a prescribed time-averaged miscoverage on a synthetic calibration sample. Under exchangeability of calibration and target observations, the bands control expected miscoverage in finite samples. In a synthetic benchmark, the proposed bands are roughly 20% the width of KOH-MCMC bands with miscoverage near the 10% target. In calibrating an age-structured SEIR model to the 2010 Malawi measles outbreak, they are again sharper and near-nominal, and the weeks they fail to cover coincide with the asymmetric epidemic decline, exposing structural inadequacy of the mechanistic model.
Abstract:Computer-model calibration is usually undertaken in service of prediction, yet neither Bayesian Kennedy-O'Hagan (KOH) calibration nor amortized simulation-based inference is built to control the predictive error of the calibrated simulator. When parameters are weakly identified or the data-model discrepancy is misspecified, the simulator-output bands induced by a nominal posterior level can be far from nominal. We propose a two-layer framework that targets predictive miscoverage directly. First, a conditional flow-matching model learns an amortized posterior for the calibration parameters, trained on simulator outputs perturbed with Gaussian-process discrepancies of varying structure. Second, conformal risk control arranges posterior draws for the target observation into nested candidate sets and selects the smallest one whose fixed-rank pointwise bands attain a prescribed time-averaged miscoverage on a synthetic calibration sample. Under exchangeability of calibration and target observations, the bands control expected miscoverage in finite samples. In a synthetic benchmark, the proposed bands are roughly 20% the width of KOH-MCMC bands with miscoverage near the 10% target. In calibrating an age-structured SEIR model to the 2010 Malawi measles outbreak, they are again sharper and near-nominal, and the weeks they fail to cover coincide with the asymmetric epidemic decline, exposing structural inadequacy of the mechanistic model.
Machine Learning (stat.ML); Machine Learning (cs.LG); Applications (stat.AP); Computation (stat.CO); Methodology (stat.ME)
arXiv:2610.31163 [stat.ML]
(or arXiv:2610.31163v1 [stat.ML] for this version)
https://doi.org/10.48550/arXiv.2610.31163 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 300 cuvinte
Subiecte conexe
Calibrating Uncertainty for Zero-Shot Adversarial CLIP
arXiv · 8 iun. 2026
Operator-Informed Gaussian Processes for Complex Helmholtz Wavefields: From Synthetic Benchmarks to In Vivo Brain Elastography
arXiv · 27 iul. 2026
TRUST-ESD: A Risk-Calibrated and Governance-Aware AI Framework for Enterprise Strategic Decision Support Under Uncertainty
arXiv · 23 iul. 2026
SAGE: Answer-Conditioned Uncertainty Targets for Verbal Uncertainty Alignment
arXiv · 11 iun. 2026