Adaptive uncertainty estimates improve image-based prediction models
Adaptive Conformal Prediction for Image Regression Models with Application to an Inertial Confinement Fusion Emulator
Machine Learning
Summary
Many computer models that analyze images only give a single answer without saying how sure they are, which can be risky in important situations. The authors designed a new approach called ACPNN that uses similar past examples to give better, local uncertainty estimates for each prediction. This method is efficient and adapts to different types of input images to tell how much to trust each output. They tested their approach on a model that predicts complex physics simulations, showing it works well to provide trustworthy uncertainty estimates.
What this means in practice
- •For simulation engineers: Provide reliable, locally adaptive uncertainty estimates for image-based simulation emulators like inertial confinement fusion models to improve decision-making.
- •For quality assurance teams in manufacturing: Use adaptive uncertainty estimates for image-based regression models to assess prediction reliability and reduce costly errors in quality control.
Authors
Carrie J. Lei-Cramer, Michael S. Jones, Laura J. Wendelberger
Abstract
Uncertainty quantification is critical in scientific machine learning, where black-box, image-based models are increasingly deployed in high-stakes settings. In many such applications, model outputs inform costly decisions, yet most methods provide only point estimates without quantifying predictive uncertainty. This challenge is compounded by the limited accessibility and interpretability of model internals, making it difficult to assess reliability across different regions of the input space. As a result, there is a growing need for methods that can provide input-dependent uncertainty estimates to guide both model development and downstream experimentation. To address this need, we propose Adaptive Conformal Prediction using Nearest Neighbors (ACPNN), an input-adaptive conformal framework for image regression. ACPNN leverages information from neighboring samples to produce locally adaptive uncertainty estimates while maintaining low computational cost. The neighborhood structure is defined using a scaled distance metric learned via a Gaussian Process with an automatic relevance determination (ARD) kernel. We demonstrate the effectiveness of ACPNN on a diffusion model for emulating inertial confinement fusion (ICF) simulations, showing that it achieves reliable and adaptive uncertainty quantification.