{"title":"利用低函数张量分解和深度解卷进行网络流量自适应异常检测","authors":"Lukas Schynol, Marius Pesavento","doi":"arxiv-2409.11529","DOIUrl":null,"url":null,"abstract":"Anomaly detection (AD) is increasingly recognized as a key component for\nensuring the resilience of future communication systems. While deep learning\nhas shown state-of-the-art AD performance, its application in critical systems\nis hindered by concerns regarding training data efficiency, domain adaptation\nand interpretability. This work considers AD in network flows using incomplete\nmeasurements, leveraging a robust tensor decomposition approach and deep\nunrolling techniques to address these challenges. We first propose a novel\nblock-successive convex approximation algorithm based on a regularized\nmodel-fitting objective where the normal flows are modeled as low-rank tensors\nand anomalies as sparse. An augmentation of the objective is introduced to\ndecrease the computational cost. We apply deep unrolling to derive a novel deep\nnetwork architecture based on our proposed algorithm, treating the\nregularization parameters as learnable weights. Inspired by Bayesian\napproaches, we extend the model architecture to perform online adaptation to\nper-flow and per-time-step statistics, improving AD performance while\nmaintaining a low parameter count and preserving the problem's permutation\nequivariances. To optimize the deep network weights for detection performance,\nwe employ a homotopy optimization approach based on an efficient approximation\nof the area under the receiver operating characteristic curve. Extensive\nexperiments on synthetic and real-world data demonstrate that our proposed deep\nnetwork architecture exhibits a high training data efficiency, outperforms\nreference methods, and adapts seamlessly to varying network topologies.","PeriodicalId":501034,"journal":{"name":"arXiv - EE - Signal Processing","volume":"4 1","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2024-09-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Adaptive Anomaly Detection in Network Flows with Low-Rank Tensor Decompositions and Deep Unrolling\",\"authors\":\"Lukas Schynol, Marius Pesavento\",\"doi\":\"arxiv-2409.11529\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Anomaly detection (AD) is increasingly recognized as a key component for\\nensuring the resilience of future communication systems. While deep learning\\nhas shown state-of-the-art AD performance, its application in critical systems\\nis hindered by concerns regarding training data efficiency, domain adaptation\\nand interpretability. This work considers AD in network flows using incomplete\\nmeasurements, leveraging a robust tensor decomposition approach and deep\\nunrolling techniques to address these challenges. We first propose a novel\\nblock-successive convex approximation algorithm based on a regularized\\nmodel-fitting objective where the normal flows are modeled as low-rank tensors\\nand anomalies as sparse. An augmentation of the objective is introduced to\\ndecrease the computational cost. We apply deep unrolling to derive a novel deep\\nnetwork architecture based on our proposed algorithm, treating the\\nregularization parameters as learnable weights. Inspired by Bayesian\\napproaches, we extend the model architecture to perform online adaptation to\\nper-flow and per-time-step statistics, improving AD performance while\\nmaintaining a low parameter count and preserving the problem's permutation\\nequivariances. To optimize the deep network weights for detection performance,\\nwe employ a homotopy optimization approach based on an efficient approximation\\nof the area under the receiver operating characteristic curve. Extensive\\nexperiments on synthetic and real-world data demonstrate that our proposed deep\\nnetwork architecture exhibits a high training data efficiency, outperforms\\nreference methods, and adapts seamlessly to varying network topologies.\",\"PeriodicalId\":501034,\"journal\":{\"name\":\"arXiv - EE - Signal Processing\",\"volume\":\"4 1\",\"pages\":\"\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-09-17\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"arXiv - EE - Signal Processing\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/arxiv-2409.11529\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"arXiv - EE - Signal Processing","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/arxiv-2409.11529","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
摘要
异常检测(AD)越来越被认为是确保未来通信系统弹性的关键组成部分。虽然深度学习已经显示出最先进的异常检测性能,但其在关键系统中的应用却受到训练数据效率、领域适应性和可解释性等问题的阻碍。本研究利用不完整的测量数据考虑网络流中的反向增量,并利用稳健的张量分解方法和深度滚动技术来应对这些挑战。我们首先提出了一种基于正则化模型拟合目标的新型块继承凸近似算法,其中正常流量被建模为低秩张量,异常流量被建模为稀疏。为了降低计算成本,我们引入了一个增强目标。我们基于所提出的算法,应用深度开卷法推导出一种新型的深度网络架构,并将其标准化参数视为可学习的权重。在贝叶斯方法的启发下,我们扩展了模型架构,以对每流和每时间步统计进行在线适应,从而提高了 AD 性能,同时保持了较低的参数数量,并保留了问题的包换方差。为了优化深度网络权重以提高检测性能,我们采用了一种同调优化方法,该方法基于对接收器工作特征曲线下面积的有效近似。在合成数据和真实世界数据上进行的广泛实验表明,我们提出的深度网络架构具有很高的训练数据效率,优于参考方法,并能无缝适应不同的网络拓扑结构。
Adaptive Anomaly Detection in Network Flows with Low-Rank Tensor Decompositions and Deep Unrolling
Anomaly detection (AD) is increasingly recognized as a key component for
ensuring the resilience of future communication systems. While deep learning
has shown state-of-the-art AD performance, its application in critical systems
is hindered by concerns regarding training data efficiency, domain adaptation
and interpretability. This work considers AD in network flows using incomplete
measurements, leveraging a robust tensor decomposition approach and deep
unrolling techniques to address these challenges. We first propose a novel
block-successive convex approximation algorithm based on a regularized
model-fitting objective where the normal flows are modeled as low-rank tensors
and anomalies as sparse. An augmentation of the objective is introduced to
decrease the computational cost. We apply deep unrolling to derive a novel deep
network architecture based on our proposed algorithm, treating the
regularization parameters as learnable weights. Inspired by Bayesian
approaches, we extend the model architecture to perform online adaptation to
per-flow and per-time-step statistics, improving AD performance while
maintaining a low parameter count and preserving the problem's permutation
equivariances. To optimize the deep network weights for detection performance,
we employ a homotopy optimization approach based on an efficient approximation
of the area under the receiver operating characteristic curve. Extensive
experiments on synthetic and real-world data demonstrate that our proposed deep
network architecture exhibits a high training data efficiency, outperforms
reference methods, and adapts seamlessly to varying network topologies.