←Back to feed
🧠 AI🔴 BearishImportance 6/10Actionable
PoiCGAN: A Targeted Poisoning Based on Feature-Label Joint Perturbation in Federated Learning
arXiv – CS AI|Tao Liu, Jiguang Lv, Dapeng Man, Weiye Xi, Yaole Li, Feiyu Zhao, Kuiming Wang, Yingchao Bian, Chen Xu, Wu Yang|
🤖AI Summary
Researchers propose PoiCGAN, a new targeted poisoning attack method for federated learning that uses feature-label joint perturbation to bypass detection mechanisms. The attack achieves 83.97% higher success rates than existing methods while maintaining model performance with less than 8.87% accuracy reduction.
Key Takeaways
- →PoiCGAN uses conditional generative adversarial networks to create stealthy poisoning attacks in federated learning systems.
- →The method achieves 83.97% higher attack success rates compared to baseline poisoning methods.
- →Poisoned models maintain high performance with less than 8.87% reduction in main task accuracy.
- →The attack is designed to bypass model anomaly detection and performance-based defense mechanisms.
- →Both poisoned samples and malicious models demonstrate high stealthiness against current detection methods.
#federated-learning#poisoning-attack#gan#cybersecurity#ai-security#machine-learning#distributed-learning#adversarial-attacks
Read Original →via arXiv – CS AI
Act on this with AI
Stay ahead of the market.
Connect your wallet to an AI agent. It reads balances, proposes swaps and bridges across 15 chains — you keep full control of your keys.
Related Articles