CARE: Adaptive Calibration for Reliable Recommendations

Modern recommender systems are typically trained offline and deployed with parameters held fixed between periodic refreshes, yet user behavior can evolve substantially during deployment. This can cause ranking utility to degrade over time and makes it difficult to provide formal guarantees about recommendation quality. We propose CARE, an adaptive calibration framework that wraps an arbitrary backbone recommender and outputs variable-size recommendation sets with finite-sample performance guarantees over interaction streams. CARE combines (i) a loss-based monitoring module that localizes behavioral changes and triggers threshold recalibration, and (ii) an online aggregation rule that promotes compact recommendation sets by dynamically reweighting candidate set predictors. We provide theoretical results establishing finite-sample guarantees for utility-based risk control and bounds on the expected set size relative to the best constituent predictor. Experiments across multiple datasets and backbone models demonstrate that CARE improves robustness and maintains compact recommendation sets while preserving the desired statistical guarantees. The code and implementation are available in https://github.com/kalpiree/CARE.

Recommendation systems are often trained on past user behavior and then deployed for some time before being updated. However, people’s interests can change during deployment, which can make recommendations less reliable. This paper introduces CARE, a method that can be added on top of existing recommender systems to monitor such changes and adjust the size of the recommendation set automatically. When the system is uncertain, CARE can show a slightly larger set of items to account for uncertainity, and when user behavior is stable, it automatically keeps the set compact. Experiments show that CARE helps maintain reliable recommendations under changing user preferences while still avoiding unnecessarily large recommendation sets.