$K^4$: Online Log Anomaly Detection Via Unsupervised Typicality Learning
This work addresses log anomaly detection for system monitoring, offering a fast and accurate solution that is incremental in improving evaluation protocols and efficiency.
The paper tackles the problem of slow and parser-dependent log anomaly detection by introducing K^4, an unsupervised framework that transforms log embeddings into compact descriptors for online detection, achieving AUROC scores of 0.995-0.999 and inference times as low as 4 μs.
Existing Log Anomaly Detection (LogAD) methods are often slow, dependent on error-prone parsing, and use unrealistic evaluation protocols. We introduce $K^4$, an unsupervised and parser-independent framework for high-performance online detection. $K^4$ transforms arbitrary log embeddings into compact four-dimensional descriptors (Precision, Recall, Density, Coverage) using efficient k-nearest neighbor (k-NN) statistics. These descriptors enable lightweight detectors to accurately score anomalies without retraining. Using a more realistic online evaluation protocol, $K^4$ sets a new state-of-the-art (AUROC: 0.995-0.999), outperforming baselines by large margins while being orders of magnitude faster, with training under 4 seconds and inference as low as 4 $μ$s.