Synergizing Large Language Models and Task-specific Models for Time Series Anomaly Detection

ArXi:2501.05675v5 Announce Type: replace In anomaly detection, methods based on large language models (LLMs) can incorporate expert knowledge by reading professional document, while task-specific small models excel at extracting normal data patterns and detecting value fluctuations from