.png){kind=icon}
返回
Multi-Task Learning for Intrusion Detection on web logs
.png){kind=icon}
PRE
AI
DOI：10.1016/j.sysarc.2017.10.011.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
In this paper, we aim to detect malicious network activities based on the analysis of web logs. Despite recent advances, classifying all malicious activities into specific types as well as identifying novel attacks are still serious issues. Various kinds of attacks have different representations. In traditional approaches, detecting each kind of attack is usually considered as an independent task. However, it is observed that different types share some common features in URL, which can be formulated as a multi-task problem. Inspired by this observation, we propose a novel Multi-Task Learning Intrusion Detection (MTLID) approach to share these common features across all types, which improves the performance of classification. Moreover, in order to detect the false negatives introduced by multi-task classification, we adopt Gaussian Mixture Model (GMM) to build the profile of normal activities, and thereby novel attacks could be further identified. We obtain a real-world dataset of web logs from different websites to demonstrate the effectiveness of MTLID. Experiment results illustrate that our proposed approach outperforms existing methods in both detection rate and false alarm rate.
Keyword:
Intrusion detection
Multi-Task learning
Web logs
Gaussian Mixture Model (GMM)
AI总结
对已上传原文的论文进行重点信息的提取，主要内容包括：简要概述、研究摘要、背景介绍、关键亮点、图文解析、展望与总结。
期刊
IF:
4.1
论文数:
3.0K
被引数:
4.2K
机构
引用论文
.png){kind=icon}
分享收藏分享收藏
查看更多内容