Reliably detecting model failures in deployment without labels

Viet Nguyen, Changjian Shui, Vijay Giri, Siddharth Arya, Amol Verma, Fahad Razak, Rahul Krishnan

Keywords: ML for Healthcare, Guardrails for AI, AI Safety, Deterioration Monitoring, Trustworthy ML

TL;DR: D-PDDM provably monitors model deterioration requiring no training data during deployment, and performs well in real-worlds datasets.

Abstract: The distribution of data changes over time; models operating in dynamic environments need retraining. But knowing when to retrain, without access to labels, is an open challenge since some, but not all shifts degrade model performance. This paper formalizes and addresses the problem of post-deployment deterioration (PDD) monitoring. We propose D3M, a practical and efficient monitoring algorithm based on the disagreement of predictive models, achieving low false positive rates under non-deteriorating shifts and provides sample complexity bounds for high true positive rates under deteriorating shifts. Empirical results on both standard benchmark and a real-world large-scale internal medicine dataset demonstrate the effectiveness of the framework and highlight its viability as an alert mechanism for high-stakes machine learning pipelines.

Primary Area: Machine learning for sciences (e.g. climate, health, life sciences, physics, social sciences)

Submission Number: 18098