Monitoring Robustness and Individual Fairness

Researchers propose runtime monitoring of black-box AI models to detect input-output robustness violations, such as adversarial or fairness issues, by observing sequences of model executions and raising alarms when similar inputs yield dissimilar outputs. They introduce the tool Clemont, which employs online FRNN algorithms and a novel binary decision diagram-based method, demonstrating effectiveness in real-time detection across standard benchmarks.