Adaptive Meta-learning-based Adversarial Training for Robust Automatic Modulation Classification

2025-01-03

AI Summary

Automatic modulation classification (AMC) models exhibit poor robustness against unknown adversarial attacks and limited adaptability in real-time online scenarios. Method: This paper introduces Model-Agnostic Meta-Learning (MAML) into AMC adversarial training for the first time, proposing a lightweight adaptive adversarial training framework that integrates Projected Gradient Descent (PGD) attack generation with few-shot gradient updates to enable zero- or few-shot real-time online defense. Contribution/Results: Experiments across multiple modulation schemes and signal-to-noise ratios demonstrate significant improvements in robust accuracy against unseen attacks—averaging a 28.7% gain over standard adversarial training—while reducing online adaptation latency by over 90%. The approach effectively overcomes key limitations of conventional methods, including weak generalization to novel attacks and inability to perform dynamic model updates.

Technology Category

• Machine Learning: Adversarial Learning & Robustness
• Multiagent Systems: Adversarial Agents
• Computer Vision: Adversarial Attacks & Robustness

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

DL-based automatic modulation classification (AMC) models are highly susceptible to adversarial attacks, where even minimal input perturbations can cause severe misclassifications. While adversarially training an AMC model based on an adversarial attack significantly increases its robustness against that attack, the AMC model will still be defenseless against other adversarial attacks. The theoretically infinite possibilities for adversarial perturbations mean that an AMC model will inevitably encounter new unseen adversarial attacks if it is ever to be deployed to a real-world communication system. Moreover, the computational limitations and challenges of obtaining new data in real-time will not allow a full training process for the AMC model to adapt to the new attack when it is online. To this end, we propose a meta-learning-based adversarial training framework for AMC models that substantially enhances robustness against unseen adversarial attacks and enables fast adaptation to these attacks using just a few new training samples, if any are available. Our results demonstrate that this training framework provides superior robustness and accuracy with much less online training time than conventional adversarial training of AMC models, making it highly efficient for real-world deployment.

Problem

• Automatic Modulation Classification
• Stability under Unseen Attacks
• Real-time Adaptation in Communication Systems

Innovation

• Meta-Learning
• Adversarial Training
• Automatic Modulation Classification