Privacy‐preserving multisource transfer learning in intrusion detection system,Transactions on Emerging Telecommunications Technologies

当前位置： X-MOL 学术 › Trans. Emerg. Telecommun. Technol. › 论文详情

Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)

Privacy‐preserving multisource transfer learning in intrusion detection system
Transactions on Emerging Telecommunications Technologies ( IF 2.5 ) Pub Date : 2020-04-14 , DOI: 10.1002/ett.3957
Mengfan Xu ₁ , Xinghua Li _{1,

2,

3} , Yunwei Wang ₁ , Bin Luo ₁ , Jingjing Guo ₁

Affiliation

The increasing scale of the network and the demand for data privacy‐preserving have brought several challenges for existing intrusion detection schemes, which presents three issues: large computational overhead, long training period, and different feature distribution which leads low model performance. The emergence of transfer learning has solved the above problems. However, the existing transfer learning‐based schemes can only operate in plaintext when different domains and clouds are untrusted entities, the privacy during data processing cannot be preserved. Therefore, this paper designs a privacy‐preserving multi‐source transfer learning intrusion detection system (IDS). Firstly, we used the Paillier homomorphic to encrypt models which trained from different source domains and uploaded to the cloud. Then, based on privacy‐preserving scheme, we first proposed a multisource transfer learning IDS based on encrypted XGBoost (E‐XGBoost). The experimental results show that the proposed scheme can successfully transfer the encryption models from multiple source domains to the target domain, and the accuracy rate can reach 93.01% in ciphertext, with no significant decrease in detection performance compared with works in plaintext. The training time of the model is significantly reduced from the traditional hour‐level to the minute‐level.

中文翻译：

入侵检测系统中保护隐私的多源转移学习

网络规模的不断扩大以及对数据隐私保护的需求给现有的入侵检测方案带来了一些挑战，这带来了三个问题：计算开销大，训练周期长以及特征分布不同导致模型性能低下。迁移学习的出现解决了上述问题。但是，当不同的域和云是不受信任的实体时，现有的基于迁移学习的方案只能以明文形式运行，无法保留数据处理期间的隐私。因此，本文设计了一种保护隐私的多源转移学习入侵检测系统（IDS）。首先，我们使用Paillier同态对从不同源域训练并上传到云的模型进行加密。然后，根据隐私保护方案，我们首先提出了一种基于加密XGBoost（E-XGBoost）的多源转移学习IDS。实验结果表明，该方案可以成功地将加密模型从多个源域转移到目标域，密文的准确率可以达到93.01％，与纯文本相比，检测性能没有明显下降。模型的训练时间从传统的小时级别显着减少到分钟级别。与纯文本作品相比，检测性能没有明显下降。模型的训练时间从传统的小时级别显着减少到分钟级别。与纯文本作品相比，检测性能没有显着下降。模型的训练时间从传统的小时级别显着减少到分钟级别。

更新日期：2020-04-14

点击分享查看原文

点击收藏

阅读更多本刊最新论文本刊介绍/投稿指南11