Threats of Adversarial Attacks in DNN-Based Modulation Recognition

Lin, Yun; Zhao, Haojun; Tu, Ya; Mao, Shiwen; Dou, Zheng

doi:10.1109/infocom41043.2020.9155389

Cited by 94 publications

(41 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Here we discuss different types of white-box attacks which are studied in the literature [12], [25]. These attack models are used in our experiments.…”

Section: Attack Modelsmentioning

confidence: 99%

A Hybrid Defense Method against Adversarial Attacks on Traffic Sign Classifiers in Autonomous Vehicles

Khan¹,

Chowdhury²,

Khan³

2022

Preprint

View full text Add to dashboard Cite

Adversarial attacks can make deep neural network (DNN) models predict incorrect output labels, such as misclassified traffic signs, for autonomous vehicle (AV) perception modules. Resilience against adversarial attacks can help AVs navigate safely on the road by avoiding misclassication of signs or objects. This DNN-based study develops a resilient traffic sign classifier for AVs that uses a hybrid defense method. We use transfer learning to retrain the Inception-V3 and Resnet-152 models as traffic sign classifiers. This method also utilizes a combination of three different strategies: random filtering, ensembling, and local feature mapping. We use the random cropping and resizing technique for random filtering, plurality voting as ensembling strategy and an optical character recognition model as a local feature mapper. This DNN-based hybrid defense method has been tested for the no attack scenario and against well-known untargeted adversarial attacks (e.g., Projected Gradient Descent or PGD, Fast Gradient Sign Method or FGSM, Momentum Iterative Method or MIM attack, and Carlini and Wagner or C&W). We find that our hybrid defense method achieves 99% average traffic sign classification accuracy for the no attack scenario and 88% average traffic sign classification accuracy for all attack scenarios. Moreover, the hybrid defense method, presented in this study, improves the accuracy for traffic sign classification compared to the traditional defense methods (i.e., JPEG filtering, feature squeezing, binary filtering, and random filtering) up to 6%, 50%, and 55% for FGSM, MIM, and PGD attacks, respectively.<br>

show abstract

“…Here we discuss different types of white-box attacks which are studied in the literature [12], [25]. These attack models are used in our experiments.…”

Section: Attack Modelsmentioning

confidence: 99%

A Hybrid Defense Method against Adversarial Attacks on Traffic Sign Classifiers in Autonomous Vehicles

Khan¹,

Chowdhury²,

Khan³

2022

Preprint

View full text Add to dashboard Cite

show abstract

“…is type of deep learning can adapt with any improvement to the hidden layers during training and the training going through backpropagation algorithm. Since DNN is good with scalable data in prediction model that uses complex data, it is considered suitable for education deep learning prediction [10,11].…”

Section: Deep Neural Network (Dnn)mentioning

confidence: 99%

Performance Prediction for Higher Education Students Using Deep Learning

Liu

2021

Complexity

View full text Add to dashboard Cite

Predicting students’ performance is very important in matters related to higher education as well as with regard to deep learning and its relationship to educational data. Prediction of students’ performance provides support in selecting courses and designing appropriate future study plans for students. In addition to predicting the performance of students, it helps teachers and managers to monitor students in order to provide support to them and to integrate the training programs to obtain the best results. One of the benefits of student’s prediction is that it reduces the official warning signs as well as expelling students because of their inefficiency. Prediction provides support to the students themselves through their choice of courses and study plans appropriate to their abilities. The proposed method used deep neural network in prediction by extracting informative data as a feature with corresponding weights. Multiple updated hidden layers are used to design neural network automatically; number of nodes and hidden layers controlled by feed forwarding and backpropagation data are produced by previous cases. The training mode is used to train the system with labeled data from dataset and the testing mode is used for evaluating the system. Mean absolute error (MAE) and root mean squared error (RMSE) with accuracy used for evolution of the proposed method. The proposed system has proven its worth in terms of efficiency through the achieved results in MAE (0.593) and RMSE (0.785) to get the best prediction.

show abstract

“…Hence, the deep learning (DL) system wrongly uses these inputs to classify input signals [10,11]. These wrong classifications of signals are not "common white noise" but a distinct attribute in the feature space that leads to incorrect model outputs [12][13][14][15][16][17][18][19][20][21][22][23].…”

Section: Literature Reviewmentioning

confidence: 99%

Analysis of performance vulnerability of MAC scheduling algorithms due to SYN flood attack in 5G NR mmWave

Saikia¹,

Majumder²

2021

IJATEE

View full text Add to dashboard Cite

Fifth-generation (5G) new radio (NR) millimetre wave (mmWave) is a kind of 5G network that operates in the 24GHz to 100GHz frequency range. It offers several opportunities as well as numerous challenges. One of the most prominent challenges that a 5G network faces is intrusion. Intrusion is possible because of existing vulnerabilities in the 5G NR mmWave network architecture. We exploited one such vulnerabil ity to create synchronise (SYN) flood intrusion into the network. The SYN flood intrusion is a denial of service (DoS) intrusion. The intruder who is involved in the SYN flood intrusion depletes the network's available resources. As a result, it denies genuine UEs/nodes access to the network's services and resources. Since this attack produces many open connections with the server, it slows down media access control (MAC) schedulers' ability to assign available channels to the user equipment. In this article, we proposed a method to exploit existing vulnerabilities of the 5G NR mmWave network to carry out SYN flood attacks. Further, we investigated the effect of the attack on the performance of the MAC schedulers, such as proportionate fair and round robin MAC schedulers. With the addition of SYN flood attack UEs/nodes, we observed that the throughput for proportional fair and round robin MAC schedulers drops dramatically.In the event of an attack, throughput drops by 2.34 % to 37.7%. However, in the event of an SYN flood attack, network delay and jitter increase. The performance of the network suffers as a result of that.

show abstract

Threats of Adversarial Attacks in DNN-Based Modulation Recognition

Cited by 94 publications

References 21 publications

A Hybrid Defense Method against Adversarial Attacks on Traffic Sign Classifiers in Autonomous Vehicles

A Hybrid Defense Method against Adversarial Attacks on Traffic Sign Classifiers in Autonomous Vehicles

Performance Prediction for Higher Education Students Using Deep Learning

Analysis of performance vulnerability of MAC scheduling algorithms due to SYN flood attack in 5G NR mmWave

Contact Info

Product

Resources

About