关键词: 异常流量检测, 对抗样本攻击, 集成学习, 多模裁决

Abstract: Currently,anomaly traffic detection models that leverage deep learning technologies are increasingly vulnerable to adversarial example attacks.Adversarial training has emerged as a potent defense mechanism against these adversarial attacks.By incorporating adversarial examples into the training process,it aims to enhance the model’s robustness,making it more resistant to similar attacks in the future.However,this approach is not without its drawbacks.While it indeed increases the model’s robustness,it also inadvertently leads to a decrease in the model’s detection accuracy.This trade-off between robustness and accuracy has become a pivotal concern in the realm of deep learning-based anomaly detection,sparking intense debate and research within the academic community.Addressing this critical issue,this paper proposes a novel framework that seeks to balance the model’s detection performance with its robustness against adversarial attacks.Drawing inspiration from ensemble learning,we construct a multi-model adversarial defense framework.This framework not only enhances the model’s adversarial robustness but also aims to improve its detection performance.By integrating proactive feature differential selection with passive adversarial training,we develop a comprehensive strategy that fortifies the model against adversarial threats while maintaining high detection accuracy.The model consists of a feature differential selection module,a detection body integration module,and a voting decision module,to address the issue that a single detection model cannot balance detection performance and robustness,and the problem of defense lagging.In the aspect of model training,we introduce a sophisticated method for constructing training data based on feature differential selection.This method involves selectively combining traffic features that exhibit significant differences,thereby creating a set of differentiated traffic example data.These examples are then used to train multiple heterogeneous detection models.This approach is designed to bolster the models’ resistance to adversarial attacks targeted at single models,presenting a more formidable challenge to attackers.Furthermore,the framework includes a novel adjudication mechanism for the detection results produced by the multiple models.Leveraging an improved heuristic population algorithm,we optimize the ensemble model’s adjudication strategy.This not only enhances the detection accuracy but also significantly increases the complexity and difficulty of generating effective adversarial examples,thereby providing an additional layer of defense.Experimental results underscore the efficacy of the proposed method.Compared to traditional single -model adversarial training approaches,the multi-model framework demonstrates a substantial improvement,with nearly a 10% increase in both accuracy and robustness.

Key words: Abnormal traffic detection, Adversarial example attack, Integrated learning, Multimode adjudication