29 August 2017 in IEEE Transactions on Industrial Electronics
Complex industrial processes may be formulated with hybrid correlations, indicating that linear and nonlinear relationships simultaneously exist among process variables, which brings great challenges for process monitoring. However, previous work did not consider the hybrid correlations and treated all the variables as a single subject in which single linear or nonlinear analysis method was employed based on prior process knowledge or some evaluation results, which may degrade the model accuracy and monitoring performance. Therefore, for complex processes with hybrid correlations, this paper proposes a linearity evaluation and variable subset partition based hierarchical modeling and monitoring method. First, linear variable subsets are separated from nonlinear subsets through an iterative variable correlation evaluation procedure. Second, hierarchical models are developed to capture linear patterns and nonlinear patterns in different levels. Third, a hierarchical monitoring strategy is proposed to monitor linear feature and nonlinear feature separately. By separating and modeling different types of variable correlations, the proposed method can explore more accurate process characteristics and thus improve the fault detection ability. Numerical examples and industrial applications are presented to illustrate its efficiency.
20 September 2021 in IEEE Transactions on Information Forensics and Security
Recent work has shown that the introduction of autonomous vehicles (AVs) in traffic could help reduce traffic jams. Deep reinforcement learning methods demonstrate good performance in complex control problems, including autonomous vehicle control, and have been used in state-of-the-art AV controllers. However, deep neural networks (DNNs) render automated driving vulnerable to machine learning-based attacks. In this work, we explore the backdooring/trojanning of DRL-based AV controllers. We develop a trigger design methodology that is based on well-established principles of traffic physics. The malicious actions include vehicle deceleration and acceleration to cause stop-and-go traffic waves to emerge (congestion attacks) or AV acceleration resulting in the AV crashing into the vehicle in front (insurance attack). We test our attack on single-lane and two-lane circuits. Our experimental results show that the backdoored model does not compromise normal operation performance, with the maximum decrease in cumulative rewards being 1%. Still, it can be maliciously activated to cause a crash or congestion when the corresponding triggers appear.
1 October 2021 in Journal of Process Control
8 March 2024 in IEEE Transactions on Artificial Intelligence
Backdoor attacks in Deep Neural Networks (DNNs) involve an attacker inserting a backdoor into the network by manipulating the training dataset, which causes misclassification of inputs that contain a specific trigger. Detecting and mitigating such attacks is challenging as only the attacker knows the trigger and target class. Our study demonstrates that the representations, i.e., the neuron activations for a given DNN, of poisoned and genuine data lie in different subspaces, which implies there exists a certain subspace where the difference of projections from different data can be manifested. To this end, we propose a method based on subspace projective clustering (SPC), which learns a subspace as well as a projection-based weight vector by solving a projection maximization program, and the optimized weight vector can be utilized in a clustering framework to infer the group of data. Based on our theoretical analysis and experimental results, we demonstrate the effectiveness of our method in defending against backdoor attacks that use different settings of poisoned samples on GTSRB, Imagenet, VGGFace2 and PubFig datasets in comparison with the state-of-the-art methods. Our algorithm can detect more than 90% of the infected classes and identify 95% of the poisoned samples.
9 August 2024 in ACM Journal on Autonomous Transportation Systems
While the advent of Deep Reinforcement Learning (DRL) has substantially improved the efficiency of Autonomous Vehicles (AVs), it makes them vulnerable to backdoor attacks that can potentially cause traffic congestion or even collisions. Backdoor functionality is typically implanted by poisoning training datasets with stealthy malicious data, designed to preserve high accuracy on legitimate inputs while inducing desired misclassifications for specific adversary-selected inputs. Existing countermeasures against backdoors predominantly concentrate on image classification, utilizing image-based properties, rendering these methods inapplicable to the regression tasks of DRL-based AV controllers that rely on continuous sensor data as inputs. In this article, we introduce the first-ever defense against backdoors on regression tasks of DRL-based models, called Backdozer . Our method systematically extracts more abstract features from representations of training data by projecting them into a specific latent subspace and segregating them into several disjoint groups based on the distribution of legitimate outputs. The key observation of Backdozer is that authentic representations for each group reside in one latent subspace, whereas the incorporation of malicious data impacts that subspace. Backdozer optimizes a sample-wise weight vector for the representations capturing the disparities in projections originating from different groups. We experimentally demonstrate that Backdozer can attain 100% accuracy in detecting backdoors. We also evaluate its effectiveness against three closely related state-of-the-art defenses.