Anomaly Detection in Networked Bandits
This work addresses the challenge of robust online learning and anomaly detection in social networks, which is crucial for preventing negative impacts from abnormal nodes, though it appears incremental as it builds on existing collaborative contextual bandit approaches.
The paper tackles the problem of detecting anomalies in networked bandits by introducing a novel algorithm that learns user preferences and detects anomalies simultaneously, achieving a proven upper bound on regret and demonstrating competitive performance against state-of-the-art methods on synthetic and real-world datasets.
The nodes' interconnections on a social network often reflect their dependencies and information-sharing behaviors. Nevertheless, abnormal nodes, which significantly deviate from most of the network concerning patterns or behaviors, can lead to grave consequences. Therefore, it is imperative to design efficient online learning algorithms that robustly learn users' preferences while simultaneously detecting anomalies. We introduce a novel bandit algorithm to address this problem. Through network knowledge, the method characterizes the users' preferences and residuals of feature information. By learning and analyzing these preferences and residuals, it develops a personalized recommendation strategy for each user and simultaneously detects anomalies. We rigorously prove an upper bound on the regret of the proposed algorithm and experimentally compare it with several state-of-the-art collaborative contextual bandit algorithms on both synthetic and real-world datasets.