Investigating a Spectral Deception Loss Metric for Training Machine Learning-based Evasion Attacks

no code implementations • 27 May 2020 • Matthew DelVecchio, Vanessa Arndorfer, William C. Headley

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.

BIG-bench Machine Learning

Effects of Forward Error Correction on Communications Aware Evasion Attacks

no code implementations • 27 May 2020 • Matthew DelVecchio, Bryse Flowers, William C. Headley

Recent work has shown the impact of adversarial machine learning on deep neural networks (DNNs) developed for Radio Frequency Machine Learning (RFML) applications.

Adversarial Attack BIG-bench Machine Learning