针对模型所有权解析的虚假声明
False Claims against Model Ownership Resolution
AI总结:
本文揭示常见模型所有权解析方案面临恶意指控者威胁,提出利用可迁移对抗样本构造虚假证据,并证明该攻击在相关方案及Amazon Rekognition API上均有效。
摘要:
深度神经网络(DNN)模型是模型所有者的宝贵知识产权,构成了竞争优势。因此,开发防止模型被盗的技术至关重要。模型所有权解析(MOR)是一类能够遏制模型盗窃的技术。MOR方案使指控者能够通过出示水印或指纹等证据,对可疑模型提出所有权主张,以证明可疑模型是从指控者拥有的源模型窃取或派生而来。现有的大多数MOR方案优先考虑针对恶意可疑方的鲁棒性,确保在可疑模型确实是被盗模型时,指控者能够胜诉。在本文中,我们表明文献中常见的MOR方案容易受到另一种同样重要但未被充分研究的鲁棒性问题影响:恶意指控者。我们展示了恶意指控者如何能够成功地对未被盗的独立可疑模型提出虚假声明。我们的核心思想是,恶意指控者可以在不被发现的情况下偏离规定的MOR流程,方法是寻找(可迁移的)对抗样本,这些样本能够成功充当针对独立可疑模型的证据。为此,我们首先概括了常见MOR方案的流程,并表明在这种概括下,防范虚假声明与防止(可迁移的)对抗样本一样具有挑战性。通过系统性的实证评估,我们表明我们的虚假声明攻击在遵循我们概括的MOR方案中总是能够成功,包括在一个真实世界模型中:Amazon的Rekognition API。
英文摘要:
Deep neural network (DNN) models are valuable intellectual property of model owners, constituting a competitive advantage. Therefore, it is crucial to develop techniques to protect against model theft. Model ownership resolution (MOR) is a class of techniques that can deter model theft. A MOR scheme enables an accuser to assert an ownership claim for a suspect model by presenting evidence, such as a watermark or fingerprint, to show that the suspect model was stolen or derived from a source model owned by the accuser. Most of the existing MOR schemes prioritize robustness against malicious suspects, ensuring that the accuser will win if the suspect model is indeed a stolen model. In this paper, we show that common MOR schemes in the literature are vulnerable to a different, equally important but insufficiently explored, robustness concern: a malicious accuser. We show how malicious accusers can successfully make false claims against independent suspect models that were not stolen. Our core idea is that a malicious accuser can deviate (without detection) from the specified MOR process by finding (transferable) adversarial examples that successfully serve as evidence against independent suspect models. To this end, we first generalize the procedures of common MOR schemes and show that, under this generalization, defending against false claims is as challenging as preventing (transferable) adversarial examples. Via systematic empirical evaluation, we show that our false claim attacks always succeed in the MOR schemes that follow our generalization, including in a real-world model: Amazon's Rekognition API.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。