arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2008.06785eess.SPcs.LG
收藏

用于安全调制分类的对抗滤波器

Adversarial Filters for Secure Modulation Classification

The University of Arizona（亚利桑那大学）

Alex Berian, Kory Staab, Noel Teku, Gregory Ditzler, Tamal Bose, Ravi Tandon

备注：
This is a placeholder to show that we are the first to do something like this. We intend to submit this work to a journal
阅读 PDF
arXiv 页面 ↗HTML 原文 ↗

摘要：

调制分类（MC）是指对无线信号的调制类别进行分类的问题。在无线通信流水线中，MC是对接收信号执行的第一个操作，并且对于可靠解码至关重要。本文考虑了安全调制分类的问题，其中发射机（爱丽丝）希望在合法接收机（鲍勃）最大限度地提高MC精度，同时在窃听者（夏娃）最小化MC精度。这项工作的贡献是设计新的对抗学习技术的安全MC。特别是，我们提出了基于对抗性过滤的算法安全MC，其中爱丽丝使用精心设计的对抗性过滤器来掩盖传输的信号，可以最大限度地提高MC准确性在鲍勃，同时最大限度地减少MC准确性在前夕。我们提出了两种基于滤波的算法，即梯度上升滤波器（GAF），和一个快速梯度滤波方法（FGFM），具有不同程度的复杂性。我们提出的基于对抗性过滤的方法显著优于加性对抗性扰动（用于传统ML社区和安全MC的其他先前工作），并且还具有其他几个理想的属性。具体地，GAF和FGFM算法是a）计算高效的（允许在Bob处的快速解码），b）功率高效的（在Alice处不需要过多的发射功率）;以及c）SNR高效的（即，即使在Bob处的低SNR值下也表现良好）。

英文摘要：

Modulation Classification (MC) refers to the problem of classifying the modulation class of a wireless signal. In the wireless communications pipeline, MC is the first operation performed on the received signal and is critical for reliable decoding. This paper considers the problem of secure modulation classification, where a transmitter (Alice) wants to maximize MC accuracy at a legitimate receiver (Bob) while minimizing MC accuracy at an eavesdropper (Eve). The contribution of this work is to design novel adversarial learning techniques for secure MC. In particular, we present adversarial filtering based algorithms for secure MC, in which Alice uses a carefully designed adversarial filter to mask the transmitted signal, that can maximize MC accuracy at Bob while minimizing MC accuracy at Eve. We present two filtering based algorithms, namely gradient ascent filter (GAF), and a fast gradient filter method (FGFM), with varying levels of complexity. Our proposed adversarial filtering based approaches significantly outperform additive adversarial perturbations (used in the traditional ML community and other prior works on secure MC) and also have several other desirable properties. In particular, GAF and FGFM algorithms are a) computational efficient (allow fast decoding at Bob), b) power-efficient (do not require excessive transmit power at Alice); and c) SNR efficient (i.e., perform well even at low SNR values at Bob).

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。

↑