New defense method shields ML models from black-box query attacks

Researchers have developed a new defense mechanism called Gradient Confusion Defense (GCD) to protect machine learning models from black-box score-based query attacks. This method, which does not require retraining the model, works by altering the probability vector returned to attackers. GCD aims to increase the complexity for adversaries attempting to craft adversarial examples by misaligning the exposed logit-gradient with the true adversarial loss gradient. Theoretical analysis and experiments on datasets like CIFAR-10 and ImageNet, using architectures such as ResNet and ViT, show that GCD effectively enhances robustness against state-of-the-art attacks without compromising predictive accuracy. AI

IMPACT Enhances security for deployed machine learning models against sophisticated adversarial attacks.

RANK_REASON Academic paper detailing a new method for mitigating adversarial attacks on machine learning models. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• AmirMahdi Sadeghzadeh
• CIFAR-10
• Gradient Confusion Defense
• ImageNet
• ResNet-18
• ResNet-50
• ViT-B/16

• paper
• safety

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →