[论文解读] Semi-WTC: A Practical Semi-supervised Framework for Attack Categorization through Weight-Task Consistency

Zihan Li, Wentao Chen+3 more|arXiv (Cornell University)|May 19, 2022

Network Security and Intrusion Detection被引用 5

一句话总结

本文提出Semi-WTC，一种用于细粒度攻击分类的半监督框架，通过权重-任务一致性（WTC）解决类别不平衡问题，并利用循环原型模块（RPM）实现高效的自监督训练。该方法在准确率上达到94.06%，相比最先进方法将训练时间减少90%。

ABSTRACT

Supervised learning has been widely used for attack categorization, requiring high-quality data and labels. However, the data is often imbalanced and it is difficult to obtain sufficient annotations. Moreover, supervised models are subject to real-world deployment issues, such as defending against unseen artificial attacks. To tackle the challenges, we propose a semi-supervised fine-grained attack categorization framework consisting of an encoder and a two-branch structure and this framework can be generalized to different supervised models. The multilayer perceptron with residual connection is used as the encoder to extract features and reduce the complexity. The Recurrent Prototype Module (RPM) is proposed to train the encoder effectively in a semi-supervised manner. To alleviate the data imbalance problem, we introduce the Weight-Task Consistency (WTC) into the iterative process of RPM by assigning larger weights to classes with fewer samples in the loss function. In addition, to cope with new attacks in real-world deployment, we propose an Active Adaption Resampling (AAR) method, which can better discover the distribution of unseen sample data and adapt the parameters of encoder. Experimental results show that our model outperforms the state-of-the-art semi-supervised attack detection methods with a 3% improvement in classification accuracy and a 90% reduction in training time.

研究动机与目标

• 解决半监督攻击分类中少数攻击类别代表性不足的数据不平衡问题。
• 通过处理未见攻击和对抗性扰动，提升模型在真实环境部署中的泛化能力与鲁棒性。
• 通过一种主动的、分布感知的样本选择机制，降低对大规模标注数据的依赖。
• 开发一种通用且可扩展的框架，兼容多种深度学习模型，适用于入侵检测任务。
• 在有限标注条件下，提升细粒度攻击检测的训练效率与分类性能。

提出的方法

• 该框架采用双分支架构，结合监督分支与无监督分支，强制预测结果的一致性。
• 使用ResBMLP（残差连接多层感知机）作为编码器，以降低复杂度并提取鲁棒特征。
• 循环原型模块（RPM）通过从伪标签化的无标签数据中迭代更新原型，实现自监督学习。
• 权重-任务一致性（WTC）在训练过程中动态为低频类别分配更高的损失权重，以缓解长尾不平衡问题。
• 主动适应重采样（AAR）选择最能代表类别分布的核心样本，其泛化能力优于基于决策边界的采样方法。
• 该框架模块化设计，可与多种监督模型集成，支持即插即用式部署。

实验结果

研究问题

• RQ1 在标注数据有限的条件下，半监督框架能否有效提升细粒度攻击分类的准确率？
• RQ2 如何在不依赖大量重加权或数据增强的前提下缓解攻击数据集中的类别不平衡问题？
• RQ3 基于分布代表性主动采样能否在入侵检测中超越传统主动学习方法？
• RQ4 该框架在真实部署中对未见攻击和对抗性样本的鲁棒性如何？
• RQ5 WTC与RPM的集成在多大程度上提升了训练效率与模型泛化能力？

主要发现

• Semi-WTC在AndMal 2020数据集上达到94.06%的分类准确率，较之前最先进方法（FARE-semisup）高出2.93个百分点。
• 与FARE-semisup相比，该模型将训练时间减少90%，仅需146秒即可完成训练。
• Semi-WTC将宏平均F1分数提升至93.72%，表明其在少数攻击类别上表现优异。
• 由于采用自监督与一致性训练机制，该框架对数据中毒攻击和对抗性样本表现出良好鲁棒性。
• 主动适应重采样（AAR）方法成功识别出具有代表性的样本，使模型泛化能力超越标准主动学习方法。
• 在AndMal 2020数据集上，该框架相较K-Means提升24%准确率，相较MixMatch提升3%准确率。