An official website of the United States government
Several attacks have been proposed against autonomous vehicles and their subsystems that are powered by machine learning (ML). Road sign recognition models are especially heavily tested under various adversarial ML attack settings, and they have proven to be vulnerable. Despite the increasing research on adversarial ML attacks against road sign recognition models, there is little to no focus on defending against these attacks. In this paper, we propose the first defense method specifically designed for autonomous vehicles to detect adversarial ML attacks targeting road sign recognition models, which is called ViLAS (Vision-Language Model for Adversarial Traffic Sign Detection). The proposed defense method is based on a custom, fast, lightweight, and salable vision-language model (VLM) and is compatible with any existing traffic sign recognition system. Thanks to the orthogonal information coming from the class label text data through the language model, ViLAS leverages image context in addition to visual data for highly effective attack detection performance. In our extensive experiments, we show that our method consistently detects various attacks against different target models with high true positive rates while satisfying very low false positive rates. When tested against four state-of-the-art attacks targeting four popular action recognition models, our proposed detector achieves an average AUC of 0.94. This result achieves a 25.3% improvement over a state-of-the-art defense method proposed for generic image attack detection, which attains an average AUC of 0.75. We also show that our custom VLM is more suitable for an autonomous vehicle compared to the popular off-the-shelf VLM and CLIP in terms of speed (4.4 vs. 9.3 milliseconds), space complexity (0.36 vs. 1.6 GB), and performance (0.94 vs. 0.43 average AUC).
more »
« less
This content will become publicly available on August 15, 2026
Robustifying ML-powered Network Classifiers with PANTS
Multiple network management tasks, from resource allocation to intrusion detection, rely on some form of ML-based network traffic classification (MNC). Despite their potential, MNCs are vulnerable to adversarial inputs, which can lead to outages, poor decision-making, and security violations, among other issues. The goal of this paper is to help network operators assess and enhance the robustness of their MNC against adversarial inputs. The most critical step for this is generating inputs that can fool the MNC while being realizable under various threat models. Compared to other ML models, finding adversarial inputs against MNCs is more challenging due to the existence of non-differentiable components e.g., traffic engineering and the need to constrain inputs to preserve semantics and ensure reliability. These factors prevent the direct use of well-established gradient-based methods developed in adversarial ML (AML). To address these challenges, we introduce PANTS, a practical white-box framework that uniquely integrates AML techniques with Satisfiability Modulo Theories (SMT) solvers to generate adversarial inputs for MNCs. We also embed PANTS into an iterative adversarial training process that enhances the robustness of MNCs against adversarial inputs. PANTS is 70% and 2x more likely in median to find adversarial inputs against target MNCs compared to state-of-the-art baselines, namely Amoeba and BAP. PANTS improves the robustness of the target MNCs by 52.7% (even against attackers outside of what is considered during robustification) without sacrificing their accuracy.
more »
« less
- PAR ID:
- 10617124
- Publisher / Repository:
- USENIX
- Date Published:
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
VehiGAN : Generative Adversarial Networks for Adversarially Robust V2X Misbehavior Detection SystemsVehicle-to-Everything (V2X) communication enables vehicles to communicate with other vehicles and roadside infrastructure, enhancing traffic management and improving road safety. However, the open and decentralized nature of V2X networks exposes them to various security threats, especially misbehaviors, necessitating a robust Misbehavior Detection System (MBDS). While Machine Learning (ML) has proved effective in different anomaly detection applications, the existing ML-based MBDSs have shown limitations in generalizing due to the dynamic nature of V2X and insufficient and imbalanced training data. Moreover, they are known to be vulnerable to adversarial ML attacks. On the other hand, Generative Adversarial Networks (GAN) possess the potential to mitigate the aforementioned issues and improve detection performance by synthesizing unseen samples of minority classes and utilizing them during their model training. Therefore, we propose the first application of GAN to design an MBDS that detects any misbehavior and ensures robustness against adversarial perturbation. In this article, we present several key contributions. First, we propose an advanced threat model for stealthy V2X misbehavior where the attacker can transmit malicious data and mask it using adversarial attacks to avoid detection by ML-based MBDS. We formulate two categories of adversarial attacks against the anomaly-based MBDS. Later, in the pursuit of a generalized and robust GAN-based MBDS, we train and evaluate a diverse set of Wasserstein GAN (WGAN) models and presentVehicularGAN(VehiGAN), an ensemble of multiple top-performing WGANs, which transcends the limitations of individual models and improves detection performance. We present a physics-guided data preprocessing technique that generates effective features for ML-based MBDS. In the evaluation, we leverage the state-of-the-art V2X attack simulation tool VASP to create a comprehensive dataset of V2X messages with diverse misbehaviors. Evaluation results show that in 20 out of 35 misbehaviors,VehiGANoutperforms the baseline and exhibits comparable detection performance in other scenarios. Particularly,VehiGANexcels in detecting advanced misbehaviors that manipulate multiple fields in V2X messages simultaneously, replicating unique maneuvers. Moreover,VehiGANprovides approximately 92% improvement in false positive rate under powerful adaptive adversarial attacks, and possesses intrinsic robustness against other adversarial attacks that target the false negative rate. Finally, we make the data and code available for reproducibility and future benchmarking, available athttps://github.com/shahriar0651/VehiGAN.more » « less
-
Adversarial machine learning (AML) research is concerned with robustness of machine learning models and algorithms to malicious tampering. Originating at the intersection between machine learning and cybersecurity, AML has come to have broader research appeal, stretching traditional notions of security to include applications of computer vision, natural language processing, and network science. In addition, the problems of strategic classification, algorithmic recourse, and counterfactual explanations have essentially the same core mathematical structure as AML, despite distinct motivations. I give a simplified overview of the central problems in AML, and then discuss both the security-motivated AML domains, and the problems above unrelated to security. These together span a number of important AI subdisciplines, but can all broadly be viewed as concerned with trustworthy AI. My goal is to clarify both the technical connections among these, as well as the substantive differences, suggesting directions for future research.more » « less
-
Machine learning-based security detection models have become prevalent in modern malware and intrusion detection systems. However, previous studies show that such models are susceptible to adversarial evasion attacks. In this type of attack, inputs (i.e., adversarial examples) are specially crafted by intelligent malicious adversaries, with the aim of being misclassified by existing state-of-the-art models (e.g., deep neural networks). Once the attackers can fool a classifier to think that a malicious input is actually benign, they can render a machine learning-based malware or intrusion detection system ineffective. Objective To help security practitioners and researchers build a more robust model against non-adaptive, white-box and non-targeted adversarial evasion attacks through the idea of ensemble model. Method We propose an approach called Omni, the main idea of which is to explore methods that create an ensemble of “unexpected models”; i.e., models whose control hyperparameters have a large distance to the hyperparameters of an adversary’s target model, with which we then make an optimized weighted ensemble prediction. Results In studies with five types of adversarial evasion attacks (FGSM, BIM, JSMA, DeepFool and Carlini-Wagner) on five security datasets (NSL-KDD, CIC-IDS-2017, CSE-CIC-IDS2018, CICAndMal2017 and the Contagio PDF dataset), we show Omni is a promising approach as a defense strategy against adversarial attacks when compared with other baseline treatments Conclusions When employing ensemble defense against adversarial evasion attacks, we suggest to create ensemble with unexpected models that are distant from the attacker’s expected model (i.e., target model) through methods such as hyperparameter optimization.more » « less
-
Research is increasingly showing the tremendous vulnerability in machine learning models to seemingly undetectable adversarial inputs. One of the current limitations in adversarial machine learning research is the incredibly time-consuming testing of novel defenses against various attacks and across multiple datasets, even with high computing power. To address this limitation, we have developed Jespipe as a new plugin-based, parallel-by-design Open MPI framework that aids in evaluating the robustness of machine learning models. The plugin-based nature of this framework enables researchers to specify any pre-training data manipulations, machine learning models, adversarial models, and analysis or visualization metrics with their input Python files. Because this framework is plugin-based, a researcher can easily incorporate model implementations using popular deep learning libraries such as PyTorch, Keras, TensorFlow, Theano, or MXNet, or adversarial robustness tools such as IBM’s Adversarial Robustness Toolbox or Foolbox. The parallelized nature of this framework also enables researchers to evaluate various learning or attack models with multiple datasets simultaneously by specifying all the models and datasets they would like to test with our XML control file template. Overall, Jespipe shows promising results by reducing latency in adversarial machine learning algorithm development and testing compared to traditional Jupyter notebook workflows.more » « less
