Transformer models with user adapters extract behavioral signals from encrypted network traffic that correlate with stress, loneliness, and sleep issues via sparse features and GEE models, outperforming handcrafted features for within-person changes.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Learning Behavioral Signals from Encrypted Smartphone Network Traffic
Transformer models with user adapters extract behavioral signals from encrypted network traffic that correlate with stress, loneliness, and sleep issues via sparse features and GEE models, outperforming handcrafted features for within-person changes.