From Packets to Patterns: Interpreting Encrypted Network Traffic as Longitudinal Behavioral Signals

2026-05-02

AI Summary

This study pioneers the use of encrypted smartphone network traffic as a longitudinal behavioral sensing modality to unobtrusively capture individual patterns related to sleep, stress, and loneliness. The authors propose an architecture built upon a Transformer backbone augmented with user-specific adapters, coupled with sparse autoencoders to extract interpretable behavioral representations. Generalized estimating equations and Mundlak decomposition are employed to disentangle between-person differences from within-person temporal dynamics. Results reveal that stress predominantly reflects stable between-person traits, loneliness manifests as within-person fluctuations, and sleep disturbances exhibit both characteristics. The learned representations significantly outperform conventional handcrafted traffic features, uncovering fine-grained behavioral signals embedded within encrypted network traffic.

Abstract

Human behavior is difficult to observe continuously at scale, yet it leaves measurable traces in everyday device use. We test whether encrypted smartphone network traffic -- a ubiquitous, always-on, passive sensing modality -- can passively capture behavioral patterns related to sleep, stress, and loneliness. We model shared behavioral structure using a transformer backbone with per-user adapters, allowing the model to represent both typical individual behavior and deviations from it. To make these representations interpretable, we apply a sparse autoencoder to extract behavioral features corresponding to distinct patterns of activity. We relate these features to sleep disturbance, stress, and loneliness using generalized estimating equations with Mundlak decomposition, separating between-person differences from within-person changes over time. We find that the three outcomes reflect distinct temporal structures: stress is primarily associated with stable between-person differences, loneliness with within-person variation, and sleep disturbance with a combination of both. Notably, these within-person dynamics are not captured by predefined network-traffic features, demonstrating the value of learned representations for longitudinal behavioral sensing. These results establish encrypted network traffic as a viable passive sensing modality, revealing interpretable behavioral dynamics -- particularly deviations from an individual's baseline -- that are not visible in raw traffic features.

Problem

• encrypted network traffic
• behavioral patterns
• longitudinal sensing
• sleep disturbance
• stress

Innovation

• encrypted network traffic
• behavioral sensing
• transformer with adapters
• sparse autoencoder
• longitudinal analysis