arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 1802.04633cs.LG
收藏

把你的弱点变成优势:通过后门对深度神经网络进行水印

Turning Your Weakness Into a Strength: Watermarking Deep Neural Networks by Backdooring

Bar-Ilan University(巴伊兰大学);Google, Inc.(谷歌公司);Facebook AI Research(Facebook人工智能研究院)

Yossi Adi, Carsten Baum, Moustapha Cisse, Benny Pinkas, Joseph Keshet

摘要:

深度神经网络最近在解决一些众所周知的挑战性问题方面取得了突破性进展,取得了很大的成功。训练这些网络在计算上是昂贵的,并且需要大量的训练数据。因此,出售这种预先训练的模型可能是一种有利可图的商业模式。不幸的是,一旦这些模型被出售,它们就很容易被复制和重新分发。为了避免这种情况,有必要建立一个跟踪机制,将模型识别为特定供应商的知识产权。在这项工作中,我们提出了一种方法,水印深度神经网络在黑盒的方式。我们的方案适用于一般的分类任务,可以很容易地与当前的学习算法相结合。我们的实验表明,这样的水印有没有明显的影响的主要任务,该模型的设计和评估的鲁棒性,我们的建议对众多的实际攻击。此外,我们提供了一个理论分析,与我们的方法以前的工作后门。

英文摘要:

Deep Neural Networks have recently gained lots of success after enabling several breakthroughs in notoriously challenging problems. Training these networks is computationally expensive and requires vast amounts of training data. Selling such pre-trained models can, therefore, be a lucrative business model. Unfortunately, once the models are sold they can be easily copied and redistributed. To avoid this, a tracking mechanism to identify models as the intellectual property of a particular vendor is necessary. In this work, we present an approach for watermarking Deep Neural Networks in a black-box way. Our scheme works for general classification tasks and can easily be combined with current learning algorithms. We show experimentally that such a watermark has no noticeable impact on the primary task that the model is designed for and evaluate the robustness of our proposal against a multitude of practical attacks. Moreover, we provide a theoretical analysis, relating our approach to previous work on backdooring.

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。

↑