arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1910.03468cs.LGcs.CVstat.ML
收藏

面向成本敏感型深度学习分类应用的定向对抗训练

Directional Adversarial Training for Cost Sensitive Deep Learning Classification Applications

University of Padova(帕多瓦大学);University of Pennsylvania(宾夕法尼亚大学)

Matteo Terzi, Gian Antonio Susto, Pratik Chaudhari

摘要:

在机器学习的许多实际应用中,不仅要提供准确的预测,而且要确保一定程度的鲁棒性,这一点至关重要。对抗训练是一种训练过程,旨在提供对预定义点周围的最坏情况扰动具有鲁棒性的模型。不幸的是,对抗训练中的一个主要问题是鲁棒性w.r.t.基于梯度的攻击总是以预测精度为代价来实现的。提出了一种新的对抗训练算法Wasserstein投影梯度下降(WPGD)。WPGD提供了一种简单的方法来获得成本敏感的鲁棒性,从而更好地控制鲁棒性-精度权衡。此外,WPGD解决了网络输出空间上的最优传输问题,并且可以有效地发现需要鲁棒性的方向,从而可以控制准确性和鲁棒性之间的方向权衡。建议的WPGD验证在这项工作中的图像识别任务与不同的基准数据集和架构。此外,类似真实世界的数据集往往是不平衡的:本文表明,在处理此类数据集时,对抗训练的性能主要受到标准准确度的影响。

英文摘要:

In many real-world applications of Machine Learning it is of paramount importance not only to provide accurate predictions, but also to ensure certain levels of robustness. Adversarial Training is a training procedure aiming at providing models that are robust to worst-case perturbations around predefined points. Unfortunately, one of the main issues in adversarial training is that robustness w.r.t. gradient-based attackers is always achieved at the cost of prediction accuracy. In this paper, a new algorithm, called Wasserstein Projected Gradient Descent (WPGD), for adversarial training is proposed. WPGD provides a simple way to obtain cost-sensitive robustness, resulting in a finer control of the robustness-accuracy trade-off. Moreover, WPGD solves an optimal transport problem on the output space of the network and it can efficiently discover directions where robustness is required, allowing to control the directional trade-off between accuracy and robustness. The proposed WPGD is validated in this work on image recognition tasks with different benchmark datasets and architectures. Moreover, real world-like datasets are often unbalanced: this paper shows that when dealing with such type of datasets, the performance of adversarial training are mainly affected in term of standard accuracy.

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。

↑