Luben M. C. Cabezas

Conformal prediction methods create prediction bands with distribution-free guarantees but do not explicitly capture epistemic uncertainty, which can lead to overconfident predictions in data-sparse regions. Although recent conformal scores have been developed to address this limitation, they are typically designed for specific tasks, such as regression or quantile regression. Moreover, they rely on particular modeling choices for epistemic uncertainty, restricting their applicability. We introduce EPICSCORE, a model-agnostic approach that enhances any conformal score by explicitly integrating epistemic uncertainty. Leveraging Bayesian techniques such as Gaussian Processes, Monte Carlo Dropout, or Bayesian Additive Regression Trees, EPICSCORE adaptively expands predictive intervals in regions with limited data while maintaining compact intervals where data is abundant. As with any conformal method, it preserves finite-sample marginal coverage. Additionally, it also achieves asymptotic conditional coverage. Experiments demonstrate its good performance compared to existing methods. Designed for compatibility with any Bayesian model, but equipped with distribution-free guarantees, EPICSCORE provides a general-purpose framework for uncertainty quantification in prediction problems.

Current experimental scientists have been increasingly relying on simulation-based inference (SBI) to invert complex nonlinear models with intractable likelihoods. However, posterior approximations obtained with SBI are often miscalibrated, causing credible regions to undercover true parameters. We develop 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 cumulative distribution function CDF-based calibration, enable finite-sample local coverage guarantees for any scoring function, including highest posterior density (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 (NPEs) using both normalizing flows and score-diffusion modelling.This article is part of the theme issue ‘Advancing uncertainty quantification in AI systems’.

Constructing valid confidence sets is a crucial task in statistical inference, yet traditional methods often face challenges when dealing with complex models or limited observed sample sizes. These challenges are frequently encountered in modern applications, such as Likelihood-Free Inference (LFI). In these settings, confidence sets may fail to maintain a confidence level close to the nominal value. In this paper, we introduce two novel methods, TRUST and TRUST++, for calibrating confidence sets to achieve distribution-free conditional coverage. These methods rely entirely on simulated data from the statistical model to perform calibration. Leveraging insights from conformal prediction techniques adapted to the statistical inference context, our methods ensure both finite-sample local coverage and asymptotic conditional coverage as the number of simulations increases, even if n is small. They effectively handle nuisance parameters and provide computationally efficient uncertainty quantification for the estimated confidence sets. This allows users to assess whether additional simulations are necessary for robust inference. Through theoretical analysis and experiments on models with both tractable and intractable likelihoods, we demonstrate that our methods outperform existing approaches, particularly in small-sample regimes. This work bridges the gap between conformal prediction and statistical inference, offering practical tools for constructing valid confidence sets in complex models.

In predictive modeling, quantifying prediction uncertainty is crucial for reliable decision-making. Traditional conformal inference methods provide marginally valid predictive regions but often produce non-adaptive intervals when naively applied to regression, potentially biasing applications. Recent advances using quantile regressors or conditional density estimators improve adaptability but are typically tied to specific prediction models, limiting their ability to quantify uncertainty around arbitrary models. Similarly, methods based on partitioning the feature space adopt sub-optimal strategies, failing to consistently measure predictive uncertainty across the feature space, especially in adversarial examples. This paper introduces a model-agnostic family of methods to calibrate prediction intervals for regression with local coverage guarantees. By leveraging regression trees and Random Forests, our approach constructs data-adaptive partitions of the feature space to approximate conditional coverage, enhancing the accuracy and scalability of prediction intervals. Our methods outperform established benchmarks on simulated and real-world datasets. They are implemented in the Python package clover, which integrates seamlessly with the scikit-learn interface for practical application.