Training Set Camouflage

Training Set Camouflage
复制标题

训练套装迷彩

DOI:
10.1007/978-3-030-01554-1_4
复制
发表时间:
2019
期刊:
International Conference on Decision and Game Theory for Security
影响因子:
--
通讯作者:
Zhu, Xiaojin
Zhu, Xiaojin
中科院分区:
--
文献类型:
--
作者:
Sen, Ayon;Alfeld, Scott;Zhang, Xuezhou;Vartanian, Ara;Ma, Yuzhe;Zhu, Xiaojin

文献摘要

参考文献

相似文献

我们在机器学习领域引入了一种隐写术,我们称之为训练集伪装。想象一下,Alice有一个关于非法机器学习分类任务的训练集。Alice希望Bob(机器学习系统)学习任务。然而,如果通信被监视,则将训练集或训练模型发送给Bob可能会引起怀疑。训练集伪装允许Alice在一个完全不同的-而且看起来是良性的-分类任务上计算第二个训练集。通过构造,发送第二套训练集不会引起怀疑。当Bob将他的标准(公共)学习算法应用于第二个训练集时,他近似地恢复了原始任务的分类器。训练集伪装是机器学习中一种新的隐写形式。我们制定训练集伪装作为一个组合的双层优化问题,并提出基于非线性规划和局部搜索的解决方案。在真实的分类任务上的实验证明了这种伪装的可行性。
We introduce a form of steganography in the domain of machine learning which we call training set camouflage. Imagine Alice has a training set on an illicit machine learning classification task. Alice wants Bob (a machine learning system) to learn the task. However, sending either the training set or the trained model to Bob can raise suspicion if the communication is monitored. Training set camouflage allows Alice to compute a second training set on a completely different – and seemingly benign – classification task. By construction, sending the second training set will not raise suspicion. When Bob applies his standard (public) learning algorithm to the second training set, he approximately recovers the classifier on the original task. Training set camouflage is a novel form of steganography in machine learning. We formulate training set camouflage as a combinatorial bilevel optimization problem and propose solvers based on nonlinear programming and local search. Experiments on real classification tasks demonstrate the feasibility of such camouflage.
DOI: 10.1007/978-0-387-74759-0_394
发表时间: 2009
期刊: --
影响因子: --
作者:
C. Floudas
通讯作者: C. Floudas
DOI: 10.1145/2020408.2020495
发表时间: 2011-08
期刊: --
影响因子: --
作者:
Michael Brückner;T. Scheffer
通讯作者: Michael Brückner;T. Scheffer
所有新闻...
DOI: --
发表时间: 1984
期刊:
影响因子: --
作者:
N. Anderson
通讯作者: N. Anderson
图像隐写术研究与文献综述
DOI: --
发表时间: 2015
期刊:
影响因子: --
作者:
Parmar Ajit Kumar Maganbhai;K. Chouhan
通讯作者: K. Chouhan
DOI: 10.1609/aaai.v32i1.11610
发表时间: 2018-01
期刊: ArXiv
影响因子: --
作者:
Xuezhou Zhang;Xiaojin Zhu;Stephen J. Wright
通讯作者: Xuezhou Zhang;Xiaojin Zhu;Stephen J. Wright