Efficient Conformal Prediction for Regression Models under Label Noise

2025-09-18

AI Summary

To address unreliable confidence interval calibration of regression models under label noise, this paper proposes a robust conformal prediction framework tailored for high-stakes applications such as medical image analysis. The method establishes a noise-agnostic threshold estimation theory and derives a quantile calibration mechanism with statistical robustness to label noise, alongside an efficient algorithm designed for continuous outputs. Crucially, it requires neither prior knowledge of noise characteristics nor explicit label cleaning steps, thereby substantially mitigating noise-induced distortions in both coverage validity and interval width. Empirical evaluation on medical imaging datasets corrupted with Gaussian label noise demonstrates that the proposed approach achieves exact marginal coverage at the nominal confidence level, while producing prediction intervals whose quality closely matches that of the noise-free baseline—outperforming existing robust conformal methods by a significant margin.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Computer Vision: Adversarial Attacks & Robustness
• Natural Language Processing: Safety and Robustness

Application Category

• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs

Abstract

In high-stakes scenarios, such as medical imaging applications, it is critical to equip the predictions of a regression model with reliable confidence intervals. Recently, Conformal Prediction (CP) has emerged as a powerful statistical framework that, based on a labeled calibration set, generates intervals that include the true labels with a pre-specified probability. In this paper, we address the problem of applying CP for regression models when the calibration set contains noisy labels. We begin by establishing a mathematically grounded procedure for estimating the noise-free CP threshold. Then, we turn it into a practical algorithm that overcomes the challenges arising from the continuous nature of the regression problem. We evaluate the proposed method on two medical imaging regression datasets with Gaussian label noise. Our method significantly outperforms the existing alternative, achieving performance close to the clean-label setting.

Problem

• Handling noisy labels in conformal prediction for regression
• Estimating noise-free confidence intervals under label corruption
• Improving reliability of regression models in medical imaging

Innovation

• Estimating noise-free CP threshold mathematically
• Developing practical algorithm for continuous regression
• Outperforming alternatives in medical imaging datasets