FreeLog combines meta-learning and adversarial domain adaptation to classify anomalies in an unlabeled target log system using labeled source logs, with reported F1 scores near 80%.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.SE 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
From Few-Label to Zero-Label: An Approach for Cross-System Log-Based Anomaly Detection with Meta-Learning
FreeLog combines meta-learning and adversarial domain adaptation to classify anomalies in an unlabeled target log system using labeled source logs, with reported F1 scores near 80%.