RIOLU automatically learns column patterns and detects pattern-violating anomalies without labels or manual parameters, reporting a 97.2% profiling F1 and beating FlashProfile and ChatGPT on four of five anomaly-detection datasets.
Data- centric systems and applications,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.SE 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Automated, Unsupervised, and Auto-parameterized Inference of Data Patterns and Anomaly Detection
RIOLU automatically learns column patterns and detects pattern-violating anomalies without labels or manual parameters, reporting a 97.2% profiling F1 and beating FlashProfile and ChatGPT on four of five anomaly-detection datasets.