A5102707204- Anomaly Detection Techniques and Applications
- Software System Performance and Reliability
- Imbalanced Data Classification Techniques
- Network Security and Intrusion Detection
- Electricity Theft Detection Techniques
- Data Quality and Management
- Big Data and Business Intelligence
Microsoft Research Asia (China)
2022-2024
Logs provide first-hand information for engineers to diagnose failures in large-scale online service systems. Log parsing, which transforms semi-structured raw log messages into structured data, is a prerequisite of automated analysis such as log-based anomaly detection and diagnosis. Almost all existing parsers follow the general idea extracting common part templates dynamic parameters. However, these parsing methods, often neglect semantic meaning messages. Furthermore, high diversity...
Cloud failure prediction (e.g., disk prediction, memory node etc.) is a crucial task for ensuring the reliability and performance of cloud systems.However, problem class imbalance poses huge challenge accurate as number healthy components (majority class) in system much larger than failed (minority class). The consequences this include biased model insufficient learning, may lack adequate information to learn characteristics associated with effectively. Moreover, current methods addressing...