研究ResNet-50在对抗攻击和恶意注入下的安全漏洞
Exploring Secure Machine Learning Through Payload Injection and FGSM Attacks on ResNet-50
- 用FGSM攻击测试模型鲁棒性，发现错误预测置信度升高
- 93.33%样本成功注入恶意载荷，图像质量无变化
- 揭示高性能模型在安全场景中的严重脆弱性
本文研究了在两种典型安全威胁下，ResNet-50图像分类模型的鲁棒性：快速梯度符号法（FGSM）对抗攻击与恶意载荷注入。模型在干净图像上的准确率为53.33%。遭受FGSM扰动后，整体准确率未变，但错误预测的置信度显著上升。同时，载荷注入方案在93.33%的测试样本中成功实施，表明攻击可在不降低视觉质量的前提下操控模型输出。这些结果凸显了即使高性能神经网络也存在严重安全隐患，亟需为安全关键应用开发更稳健的防御机制。
原文摘要 · Abstract (English)
This paper investigates the resilience of a ResNet-50 image classification model under two prominent security threats: Fast Gradient Sign Method (FGSM) adversarial attacks and malicious payload injection. Initially, the model attains a 53.33% accuracy on clean images. When subjected to FGSM perturbations, its overall accuracy remains unchanged; however, the model's confidence in incorrect predictions notably increases. Concurrently, a payload injection scheme is successfully executed in 93.33% of the tested samples, revealing how stealthy attacks can manipulate model predictions without degrading visual quality. These findings underscore the vulnerability of even high-performing neural networks and highlight the urgency of developing more robust defense mechanisms for security-critical applications.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。