Robust Uncertainty Estimation under Distribution Shift via Difference Reconstruction

2026-01-27

AI Summary

Deep learning models often fail to reliably estimate uncertainty under distributional shift, limiting their deployment in high-stakes applications such as medical imaging. To address this challenge, this work proposes Difference-based Reconstruction Uncertainty Estimation (DRUE), a method that reconstructs the input from two distinct intermediate layers and quantifies the discrepancy between their outputs. By avoiding direct comparison between the original and reconstructed images, DRUE mitigates information loss and reduces sensitivity to superficial details. Integrated within an out-of-distribution (OOD) detection framework, the approach demonstrates superior performance on multiple OOD datasets in the task of glaucoma detection, achieving consistent improvements in both AUC and AUPR metrics, thereby exhibiting enhanced robustness and reliability.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Computer Vision: Diffusion Models for Vision
• Reasoning under Uncertainty: Uncertainty Representations

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Estimating uncertainty in deep learning models is critical for reliable decision-making in high-stakes applications such as medical imaging. Prior research has established that the difference between an input sample and its reconstructed version produced by an auxiliary model can serve as a useful proxy for uncertainty. However, directly comparing reconstructions with the original input is degraded by information loss and sensitivity to superficial details, which limits its effectiveness. In this work, we propose Difference Reconstruction Uncertainty Estimation (DRUE), a method that mitigates this limitation by reconstructing inputs from two intermediate layers and measuring the discrepancy between their outputs as the uncertainty score. To evaluate uncertainty estimation in practice, we follow the widely used out-of-distribution (OOD) detection paradigm, where in-distribution (ID) training data are compared against datasets with increasing domain shift. Using glaucoma detection as the ID task, we demonstrate that DRUE consistently achieves superior AUC and AUPR across multiple OOD datasets, highlighting its robustness and reliability under distribution shift. This work provides a principled and effective framework for enhancing model reliability in uncertain environments.

Problem

• uncertainty estimation
• distribution shift
• out-of-distribution detection
• deep learning
• model reliability

Innovation

• Difference Reconstruction
• Uncertainty Estimation
• Distribution Shift
• Out-of-Distribution Detection
• Deep Learning Robustness