Robustness Testing of Data and Knowledge Driven Anomaly Detection in Cyber-Physical Systems
Robustness Testing of Data and Knowledge Driven Anomaly Detection in Cyber-Physical Systems
复制标题
网络物理系统中数据和知识驱动的异常检测的鲁棒性测试
DOI：
10.48550/arxiv.2204.09183
复制
发表时间：
2022
期刊：
影响因子：
--
通讯作者：
H. Alemzadeh
中科院分区：
文献类型：
--
作者：
Xugui Zhou;Maxfield Kouzel;H. Alemzadeh
网络物理系统（CPS）日益复杂，在确保安全性方面面临挑战，这导致越来越多地使用深度学习方法来进行准确和可扩展的异常检测。然而，机器学习（ML）模型在预测意外数据时往往性能低下，并且容易受到意外或恶意干扰。虽然深度学习模型的鲁棒性测试在图像分类和语音识别等应用中得到了广泛的探索，但CPS中ML驱动的安全监控却很少受到关注。本文提出了初步结果评估的鲁棒性ML为基础的异常检测方法在安全关键的CPS对两种类型的意外和恶意的输入扰动，使用基于高斯的噪声模型和快速梯度符号法（FGSM）。我们检验了整合领域知识（例如，不安全的系统行为）可以提高异常检测的鲁棒性，而不会牺牲准确性和透明度。针对糖尿病管理的人工胰腺系统（APS）的两个案例研究的实验结果表明，使用领域知识训练的基于ML的安全监控器可以平均减少高达54.2%的鲁棒性错误，并保持平均F1分数高，同时提高透明度。
The growing complexity of Cyber-Physical Systems (CPS) and challenges in ensuring safety and security have led to the increasing use of deep learning methods for accurate and scalable anomaly detection. However, machine learning (ML) models often suffer from low performance in predicting unexpected data and are vulnerable to accidental or malicious perturbations. Although robustness testing of deep learning models has been extensively explored in applications such as image classification and speech recognition, less attention has been paid to ML-driven safety monitoring in CPS. This paper presents the preliminary results on evaluating the robustness of ML-based anomaly detection methods in safety-critical CPS against two types of accidental and malicious input perturbations, generated using a Gaussian-based noise model and the Fast Gradient Sign Method (FGSM). We test the hypothesis of whether integrating the domain knowledge (e.g., on unsafe system behavior) with the ML models can improve the robustness of anomaly detection without sacrificing accuracy and transparency. Experimental results with two case studies of Artificial Pancreas Systems (APS) for diabetes management show that ML-based safety monitors trained with domain knowledge can reduce on average up to 54.2% of robustness error and keep the average F1 scores high while improving transparency.
登录
查看更多内容
DOI：
10.1109/ro-man50785.2021.9515358
发表时间：
2021-08
期刊：
2021 30th IEEE International Conference on Robot & Human Interactive Communication (RO-MAN)
影响因子：
--
作者：
Md Masudur Rahman;R. Voyles;J. Wachs;Yexiang Xue
通讯作者：
Md Masudur Rahman;R. Voyles;J. Wachs;Yexiang Xue
影响因子：
35.6
作者：
Olowononi, Felix O.;Rawat, Danda B.;Liu, Chunmei
通讯作者：
Liu, Chunmei
DOI：
--
发表时间：
2017-11
期刊：
--
影响因子：
--
作者：
Jingyi Xu;Zilu Zhang;Tal Friedman;Yitao Liang;Guy Van den Broeck
通讯作者：
Jingyi Xu;Zilu Zhang;Tal Friedman;Yitao Liang;Guy Van den Broeck
DOI：
10.1109/dsn48063.2020.00054
发表时间：
2020-05
期刊：
2020 50th Annual IEEE/IFIP International Conference on Dependable Systems and Networks (DSN)
影响因子：
--
作者：
M. S. Yasar;H. Alemzadeh
通讯作者：
M. S. Yasar;H. Alemzadeh
用于人工胰腺系统危险预测的上下文感知监视器的数据驱动设计
DOI：
10.1109/dsn48987.2021.00058
发表时间：
2021
期刊：
2021 51st Annual IEEE/IFIP International Conference on Dependable Systems and Networks (DSN
影响因子：
--
作者：
Zhou, Xugui;Ahmed, Bulbul;Aylor, James H.;Asare, Philip;Alemzadeh, Homa
通讯作者：
Alemzadeh, Homa