LogBERT: Log Anomaly Detection via BERT

被引:152
作者
Guo, Haixuan [1 ]
Yuan, Shuhan [1 ]
Wu, Xintao [2 ]
机构
[1] Utah State Univ, Logan, UT 84322 USA
[2] Univ Arkansas, Fayetteville, AR 72701 USA
来源
2021 INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS (IJCNN) | 2021年
基金
美国国家科学基金会;
关键词
D O I
10.1109/IJCNN52387.2021.9534113
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Detecting anomalous events in online computer systems is crucial to protect the systems from malicious attacks or malfunctions. System logs, which record detailed information of computational events, are widely used for system status analysis. In this paper, we propose LogBERT, a self-supervised framework for log anomaly detection based on Bidirectional Encoder Representations from Transformers (BERT). LogBERT learns the patterns of normal log sequences by two novel self-supervised training tasks, masked log message prediction and volume of hypersphere minimization. After training, LogBERT is able to capture the patterns of normal log sequences and further detect anomalies where the underlying patterns deviate from expected patterns. The experimental results on three log datasets show that LogBERT outperforms state-of-the-art approaches for anomaly detection.
引用
收藏
页数:8
相关论文
共 24 条
[1]  
[Anonymous], 2008, WSDM
[2]  
[Anonymous], 2019, IJCAI
[3]  
[Anonymous], 2009, SER SOSP 09, DOI DOI 10.1145/1629575.1629587
[4]  
[Anonymous], 2019, SER CCS 19, DOI DOI 10.1145/3319535.3363224
[5]  
[Anonymous], 2017, SER CCS 17, DOI DOI 10.1145/3133956.3134015
[6]   Event Logs for the Analysis of Software Failures: A Rule-Based Approach [J].
Cinque, Marcello ;
Cotroneo, Domenico ;
Pecchia, Antonio .
IEEE TRANSACTIONS ON SOFTWARE ENGINEERING, 2013, 39 (06) :806-821
[7]  
Devlin J, 2019, 2019 CONFERENCE OF THE NORTH AMERICAN CHAPTER OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS: HUMAN LANGUAGE TECHNOLOGIES (NAACL HLT 2019), VOL. 1, P4171
[8]   Drain: An Online Log Parsing Approach with Fixed Depth Tree [J].
He, Pinjia ;
Zhu, Jieming ;
Zheng, Zibin ;
Lyu, Michael R. .
2017 IEEE 24TH INTERNATIONAL CONFERENCE ON WEB SERVICES (ICWS 2017), 2017, :33-40
[9]  
He S., 2020, Loghub: A large collection of system log datasets towards automated log analytics
[10]   Experience Report: System Log Analysis for Anomaly Detection [J].
He, Shilin ;
Zhu, Jieming ;
He, Pinjia ;
Lyu, Michael R. .
2016 IEEE 27TH INTERNATIONAL SYMPOSIUM ON SOFTWARE RELIABILITY ENGINEERING (ISSRE), 2016, :207-218