Adversarial Robust Modulation Recognition Guided by Attention Mechanisms
Deep neural networks have demonstrated considerable effectiveness in recognizing complex communications signals through their applications in the tasks of automatic modulation recognition. However, the resilience of these networks is undermined by the introduction of carefully designed adversarial e...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2025-01-01
|
| Series: | IEEE Open Journal of Signal Processing |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10829960/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832583136078200832 |
|---|---|
| author | Quanhai Zhan Xiongwei Zhang Meng Sun Lei Song Zhenji Zhou |
| author_facet | Quanhai Zhan Xiongwei Zhang Meng Sun Lei Song Zhenji Zhou |
| author_sort | Quanhai Zhan |
| collection | DOAJ |
| description | Deep neural networks have demonstrated considerable effectiveness in recognizing complex communications signals through their applications in the tasks of automatic modulation recognition. However, the resilience of these networks is undermined by the introduction of carefully designed adversarial examples that compromise the reliability of the decision processes. In order to address this issue, an Attention-Guided Automatic Modulation Recognition (AG-AMR) method is proposed in this paper. The method introduces an optimized attention mechanism within the Transformer framework, where signal features are extracted and filtered based on the weights of the attention module during the training process, which makes the model to focus on key features for the task. Furthermore, by removing features of low importance where adversarial perturbations may appear, the proposed method mitigates the negative impacts of adversarial perturbations on modulation classification, thereby it improves both accuracy and robustness. Experimental results on benchmark datasets show that AG-AMR obtains a high level of accuracy on modulation recognition and exhibits significant robustness. Furthermore, when working together with adversarial training, it is shown that AG-AMR effectively resists several existing adversarial attacks, which thus further validates its effectiveness on defending against adversarial sample attacks. |
| format | Article |
| id | doaj-art-095d7d2181b34863905ab19a60d940c7 |
| institution | Kabale University |
| issn | 2644-1322 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Open Journal of Signal Processing |
| spelling | doaj-art-095d7d2181b34863905ab19a60d940c72025-01-29T00:01:34ZengIEEEIEEE Open Journal of Signal Processing2644-13222025-01-016172910.1109/OJSP.2025.352657710829960Adversarial Robust Modulation Recognition Guided by Attention MechanismsQuanhai Zhan0Xiongwei Zhang1https://orcid.org/0000-0003-4890-0668Meng Sun2https://orcid.org/0000-0002-7435-3752Lei Song3Zhenji Zhou4Lab of Intelligent Information Processing, Army Engineering University, Nanjing, ChinaLab of Intelligent Information Processing, Army Engineering University, Nanjing, ChinaLab of Intelligent Information Processing, Army Engineering University, Nanjing, ChinaLab of Intelligent Information Processing, Army Engineering University, Nanjing, ChinaLab of Intelligent Information Processing, Army Engineering University, Nanjing, ChinaDeep neural networks have demonstrated considerable effectiveness in recognizing complex communications signals through their applications in the tasks of automatic modulation recognition. However, the resilience of these networks is undermined by the introduction of carefully designed adversarial examples that compromise the reliability of the decision processes. In order to address this issue, an Attention-Guided Automatic Modulation Recognition (AG-AMR) method is proposed in this paper. The method introduces an optimized attention mechanism within the Transformer framework, where signal features are extracted and filtered based on the weights of the attention module during the training process, which makes the model to focus on key features for the task. Furthermore, by removing features of low importance where adversarial perturbations may appear, the proposed method mitigates the negative impacts of adversarial perturbations on modulation classification, thereby it improves both accuracy and robustness. Experimental results on benchmark datasets show that AG-AMR obtains a high level of accuracy on modulation recognition and exhibits significant robustness. Furthermore, when working together with adversarial training, it is shown that AG-AMR effectively resists several existing adversarial attacks, which thus further validates its effectiveness on defending against adversarial sample attacks.https://ieeexplore.ieee.org/document/10829960/Adversarial attacksadversarial trainingattention mechanismsautomatic modulation recognition |
| spellingShingle | Quanhai Zhan Xiongwei Zhang Meng Sun Lei Song Zhenji Zhou Adversarial Robust Modulation Recognition Guided by Attention Mechanisms IEEE Open Journal of Signal Processing Adversarial attacks adversarial training attention mechanisms automatic modulation recognition |
| title | Adversarial Robust Modulation Recognition Guided by Attention Mechanisms |
| title_full | Adversarial Robust Modulation Recognition Guided by Attention Mechanisms |
| title_fullStr | Adversarial Robust Modulation Recognition Guided by Attention Mechanisms |
| title_full_unstemmed | Adversarial Robust Modulation Recognition Guided by Attention Mechanisms |
| title_short | Adversarial Robust Modulation Recognition Guided by Attention Mechanisms |
| title_sort | adversarial robust modulation recognition guided by attention mechanisms |
| topic | Adversarial attacks adversarial training attention mechanisms automatic modulation recognition |
| url | https://ieeexplore.ieee.org/document/10829960/ |
| work_keys_str_mv | AT quanhaizhan adversarialrobustmodulationrecognitionguidedbyattentionmechanisms AT xiongweizhang adversarialrobustmodulationrecognitionguidedbyattentionmechanisms AT mengsun adversarialrobustmodulationrecognitionguidedbyattentionmechanisms AT leisong adversarialrobustmodulationrecognitionguidedbyattentionmechanisms AT zhenjizhou adversarialrobustmodulationrecognitionguidedbyattentionmechanisms |