DeepLog: Anomaly Detection and Diagnosis from System Logs through Deep Learning

Explore this paper's citation graph

Summary

DeepLog, a deep neural network model utilizing Long Short-Term Memory (LSTM), is proposed, to model a system log as a natural language sequence, which allows DeepLog to automatically learn log patterns from normal execution, and detect anomalies when log patterns deviate from the model trained from log data under normal execution.

Type
article
Published
2017-10-30
Cited by
2,019
References
46

Keywords

Computer science, Anomaly detection, Debugging, Anomaly (physics), Data mining

References

Cited by

Related papers