Temporal Logical Attention Network for Log-Based Anomaly Detection in Distributed Systems
2024

Temporal Logical Attention Network for Log-Based Anomaly Detection in Distributed Systems

Sample size: 100000 publication 10 minutes Evidence: high

Author Information

Author(s): Liu Yang, Ren Shaochen, Wang Xuran, Zhou Mengjie

Primary Institution: Worcester Polytechnic Institute

Hypothesis

Can a novel deep learning framework effectively detect anomalies in distributed system logs by integrating temporal sequence modeling with logical dependency analysis?

Conclusion

The TLAN framework significantly improves anomaly detection in distributed systems, achieving a 9.4% increase in F1-score and a 15.3% reduction in false alarms.

Supporting Evidence

  • TLAN outperforms existing methods by achieving a 9.4% improvement in F1-score.
  • Extensive experiments on a large-scale synthetic dataset validate the effectiveness of TLAN.
  • TLAN reduces false alarms by 15.3% while maintaining low latency in real-time detection.

Takeaway

This study created a smart system that can find problems in computer networks by looking at logs, which are like diaries of what the computers do. It works better than older methods.

Methodology

The study used a novel deep learning framework called TLAN that combines temporal sequence modeling and logical dependency analysis for anomaly detection in distributed system logs.

Limitations

The model's performance in detecting network-related anomalies suggests room for improvement in modeling distributed communication patterns.

Statistical Information

Statistical Significance

p<0.05

Digital Object Identifier (DOI)

10.3390/s24247949

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication