Uncertainty-Aware Variational Information Pursuit for Interpretable Medical Image Analysis

2025-06-20

AI Summary

Existing methods for explainable AI in medical imaging struggle to balance interpretability and reliability, primarily by neglecting instance-level uncertainty—leading to untrustworthy explanations. To address this, we propose an uncertainty-aware end-to-end interpretable learning paradigm. For the first time, our approach jointly models epistemic and aleatoric uncertainty within the Variational Information Bottleneck (V-IB) framework, integrating Bayesian deep learning, concept-level attention, and differentiable concept discovery. This enables precise, uncertainty-quantified attribution while preserving explanation conciseness and clinical interpretability. Evaluated on four benchmark datasets—PH2, Derm7pt, BrEaST, and SkinCon—our method achieves a 3.2% average AUC improvement, reduces explanation length by 20%, and maintains full information retention. These results significantly enhance model trustworthiness and clinical applicability.

Technology Category

• Computer Vision: Interpretability, Explainability, and Transparency
• Machine Learning: Calibration & Uncertainty Quantification
• Humans and AI: Explainable AI (XAI) for Human Understanding

Application Category

• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

In medical imaging, AI decision-support systems must balance accuracy and interpretability to build user trust and support effective clinical decision-making. Recently, Variational Information Pursuit (V-IP) and its variants have emerged as interpretable-by-design modeling techniques, aiming to explain AI decisions in terms of human-understandable, clinically relevant concepts. However, existing V-IP methods overlook instance-level uncertainties in query-answer generation, which can arise from model limitations (epistemic uncertainty) or variability in expert responses (aleatoric uncertainty). This paper introduces Uncertainty-Aware V-IP (UAV-IP), a novel framework that integrates uncertainty quantification into the V-IP process. We evaluate UAV-IP across four medical imaging datasets, PH2, Derm7pt, BrEaST, and SkinCon, demonstrating an average AUC improvement of approximately 3.2% while generating 20% more concise explanations compared to baseline V-IP, without sacrificing informativeness. These findings highlight the importance of uncertainty-aware reasoning in interpretable by design models for robust and reliable medical decision-making.

Problem

• Balancing accuracy and interpretability in medical AI systems
• Addressing uncertainties in interpretable medical image analysis
• Improving reliability of AI decisions in clinical settings

Innovation

• Integrates uncertainty quantification into V-IP
• Improves AUC by 3.2% on medical datasets
• Generates 20% more concise explanations