Epistemic Uncertainty-Aware Defect Detection for Quality Control in Medical Device Manufacturing

2026-10-06

AI Summary

This study addresses the reliability limitations caused by model uncertainty in medical device defect detection by proposing a theoretically grounded, uncertainty-aware selective prediction framework. The method integrates knowledge graphs to process heterogeneous manufacturing data and designs a rejection mechanism based on epistemic uncertainty, enabling an explicitly controllable trade-off between prediction coverage and accuracy. Evaluated on real-world FDA datasets, the framework demonstrates that a 10% rejection rate reduces classification errors by 48%, while an aggressive rejection strategy achieves near-perfect accuracy. These results establish the proposed approach as a robust solution that effectively balances safety and reliability for high-risk quality inspection scenarios.

Abstract

Objective: We investigate whether accounting for epistemic uncertainty can improve the reliability of automated defect detection in medical device manufacturing. Methods: We consider a machine learning framework that operates on heterogeneous manufacturing and device-report data represented with Knowledge Graphs. To mitigate errors arising from uncertainty in the decision model, we analyze a principled rejection strategy to abstain from predictions whose estimated epistemic uncertainty exceeds a specified threshold. We evaluate the approach using standard synthetic benchmarks and real-world medical device report data. Results: The theoretical results establish the validity of the method characterizing the regimes under which it is expected to be effective. Empirically, the rejection strategy enables explicit control of coverage, that is, the proportion of samples for which the model issues predictions, while improving performance on the retained samples. On 266,170 real-world FDA MAUDE device reports, a 10% abstention rate reduces classification error by 48%, and more aggressive rejection (approximately 70% coverage) yields near-perfect accuracy on the retained samples. On standard synthetic manufacturing benchmarks, abstaining on 9% of the decisions, our approach reduces the risk up to 63% compared with the standard no-abstention approach. Conclusions: Abstaining from predictions with high epistemic uncertainty can provide a practical tool for controlling the reliability of machine learning-based defect detection, especially in high-stakes medical device manufacturing applications. Significance: Uncertainty-aware defect detection may support safer and more reliable quality assurance in medical device manufacturing by identifying cases that require additional inspection rather than issuing potentially harmful predictions.

Problem

• Epistemic Uncertainty
• Defect Detection
• Medical Device Manufacturing
• Quality Control
• Reliability

Innovation

• Epistemic Uncertainty
• Defect Detection
• Knowledge Graphs
• Rejection Strategy
• Medical Device Manufacturing