FreeUp decomposes encrypted traffic into frequency-specific branches with dedicated training and uses uncertainty-inspired fusion to produce more reliable anomaly scores than standard reconstruction methods.
GLAD: towards better reconstruction with global and local adaptive diffusion models for unsupervised anomaly detection
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CR 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Decompose to Understand, Fuse to Detect: Frequency-Decoupled Anomaly Detection for Encrypted Network Traffic
FreeUp decomposes encrypted traffic into frequency-specific branches with dedicated training and uses uncertainty-inspired fusion to produce more reliable anomaly scores than standard reconstruction methods.