GESR detects anomalous hosts and communications by forcing a model to predict edge semantics from graph topology alone under benign-only training, yielding ROC-AUC 0.9753 on CICIDS2017 at tight false-positive rates.
In: Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD), pp
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
-
GESR: Graph-Based Edge Semantic Reconstruction for Stealthy Communication Detection with Benign-Only Training
GESR detects anomalous hosts and communications by forcing a model to predict edge semantics from graph topology alone under benign-only training, yielding ROC-AUC 0.9753 on CICIDS2017 at tight false-positive rates.