A Representation Learning Approach to Feature Drift Detection in Wireless Networks

TL;DR

This paper introduces ALERT, a novel method combining representation learning, statistical testing, and utility assessment to detect feature drift in wireless networks, improving model reliability in dynamic environments.

Contribution

The paper presents ALERT, a new approach that effectively detects feature distribution changes in wireless networks, outperforming ten existing drift detection methods.

Findings

ALERT outperforms ten standard drift detection methods.

Effective detection in wireless fingerprinting and link anomaly detection.

Utilizes MLP, KS test, and PSI for robust drift detection.

Abstract

AI is foreseen to be a centerpiece in next generation wireless networks enabling enabling ubiquitous communication as well as new services. However, in real deployment, feature distribution changes may degrade the performance of AI models and lead to undesired behaviors. To counter for undetected model degradation, we propose ALERT; a method that can detect feature distribution changes and trigger model re-training that works well on two wireless network use cases: wireless fingerprinting and link anomaly detection. ALERT includes three components: representation learning, statistical testing and utility assessment. We rely on MLP for designing the representation learning component, on Kolmogorov-Smirnov and Population Stability Index tests for designing the statistical testing and a new function for utility assessment. We show the superiority of the proposed method against ten standard drift detection methods available in the literature on two wireless network use cases.