Communications Aware Adversarial Residual Networks for over the Air Evasion Attacks

Flowers, Bryse; Buehrer, R. Michael; Headley, William C.

doi:10.1109/milcom47813.2019.9020716

Cited by 22 publications

(30 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…That is, they yield a high BER. However, the ability to formally define a perceptible perturbation as BER has allowed recent works to create differentiable versions of the receive chain, that allow for the BER to be directly incorporated into the loss function of an adversarial attack [84], [86], yielding more sophisticated and effective threats.…”

Section: ) Defining Perceptible Perturbations In Wirelessmentioning

confidence: 99%

An RFML Ecosystem: Considerations for the Application of Deep Learning to Spectrum Situational Awareness

Wong

Clark

Flowers

et al. 2021

IEEE Open J. Commun. Soc.

Self Cite

View full text Add to dashboard Cite

While deep learning (DL) technologies are now pervasive in state-of-the-art Computer Vision (CV) and Natural Language Processing (NLP) applications, only in recent years have these technologies started to sufficiently mature in applications related to wireless communications, a field loosely termed Radio Frequency Machine Learning (RFML). In particular, recent research has shown DL to be an enabling technology for Cognitive Radio (CR) applications as well as a useful tool for supplementing expertly defined algorithms for spectrum awareness applications such as signal detection, estimation, and classification. A major driver for the usage of RFML is that little, to no, a priori knowledge of the intended spectral environment is required, given that there is an abundance of representative raw Radio Frequency (RF) data to facilitate training and evaluation. However, in addition to this fundamental need for sufficient data, there are other key considerations, such as trust, security, and hardware requirements, that must be taken into account before deploying RFML systems in real-world wireless communication applications that largely go unaddressed in the current literature. This paper examines the prior works related to these major research considerations, with focus on the dependencies between them and factors unique to the RFML space.

show abstract

Section: ) Defining Perceptible Perturbations In Wirelessmentioning

confidence: 99%

An RFML Ecosystem: Considerations for the Application of Deep Learning to Spectrum Situational Awareness

Wong

Clark

Flowers

et al. 2021

IEEE Open J. Commun. Soc.

Self Cite

View full text Add to dashboard Cite

show abstract

“…The adversarial scenario considered in this work is illustrated by Figure 1. As previously mentioned, balancing the two conflicting goals of evasion of an eavesdropper and successful communication by an intended receiver is difficult and has been examined in previous work [4,5,10,15]. In this work, the idea of successful perception of the signal at the receiver is driven using metrics such as bit error rate (BER) that indicate the success of the communication.…”

Section: Introductionmentioning

confidence: 99%

“…In addition, this work presents a novel form of perception to be considered alongside BER, that of spectral integrity. The previous works in this area have shown that adversarial perturbations naturally tend to manifest out of the main lobe of the signal and lead to adversarial signals that do not hold well the same spectral shape as the original signal [5]. This change in the spectral shape of the signal poses a problem to the success of the attack as the eavesdropper could leverage preprocessing stages to reduce the impact of the perturbation, such as with a filter, and potentially can lead to increased likelihood of detection that an attack is taking place since the spectral shape does not appear benign.…”

Section: Introductionmentioning

confidence: 99%

“…This approaches offers improvement over previous methods where the perturbation was simply power limited in the hope that this would lead to decreased BER. Flowers et al improved upon these prior works through the development of a so-called "communications-aware" attack [5].…”

Section: Introductionmentioning

confidence: 99%

“…These three losses are each weighted and summed together to guide the ARNs learning process. The work of [4] expanded on the communications aware architecture first introduced in [5] in order to better utilize forward error correction (FEC), but it was found that the changes to the loss functions and transmitter architecture provided improvements beyond just utilizing FEC. The training framework presented in [4], and illustrated in Figure 2, serves as the foundation for the work presented in this paper.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Investigating a spectral deception loss metric for training machine learning-based evasion attacks

DelVecchio

Arndorfer

Headley

2020

Proceedings of the 2nd ACM Workshop on Wireless Security and Machine Learning

Self Cite

View full text Add to dashboard Cite

Adversarial evasion attacks have been very successful in causing poor performance in a wide variety of machine learning applications. One such application is radio frequency spectrum sensing. While evasion attacks have proven particularly successful in this area, they have done so at the detriment of the signal's intended purpose. More specifically, for real-world applications of interest, the resulting perturbed signal that is transmitted to evade an eavesdropper must not deviate far from the original signal, less the intended information is destroyed. Recent work by the authors and others has demonstrated an attack framework that allows for intelligent balancing between these conflicting goals of evasion and communication. However, while these methodologies consider creating adversarial signals that minimize communications degradation, they have been shown to do so at the expense of the spectral shape of the signal. This opens the adversarial signal up to defenses at the eavesdropper such as filtering, which could render the attack ineffective. To remedy this, this work introduces a new spectral deception loss metric that can be implemented during the training process to force the spectral shape to be more in-line with the original signal. As an initial proof of concept, a variety of methods are presented that provide a starting point for this proposed loss. Through performance analysis, it is shown that these techniques are effective in controlling the shape of the adversarial signal. CCS CONCEPTS• Computing methodologies → Neural networks; Supervised learning by classification.

show abstract

Adversarial Machine Learning for 5G Communications Security

Sagduyu

Erpek

Shi

2021

Game Theory and Machine Learning for Cyber Security

View full text Add to dashboard Cite

Communications Aware Adversarial Residual Networks for over the Air Evasion Attacks

Cited by 22 publications

References 10 publications

An RFML Ecosystem: Considerations for the Application of Deep Learning to Spectrum Situational Awareness

An RFML Ecosystem: Considerations for the Application of Deep Learning to Spectrum Situational Awareness

Investigating a spectral deception loss metric for training machine learning-based evasion attacks

Adversarial Machine Learning for 5G Communications Security

Contact Info

Product

Resources

About