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 1905.08232cs.LGcs.CRcs.CVstat.ML
收藏

对抗鲁棒的迁移学习

Adversarially robust transfer learning

University of Maryland(马里兰大学);Cornell University(康奈尔大学)

Ali Shafahi, Parsa Saadatpanah, Chen Zhu, Amin Ghiasi, Christoph Studer, David Jacobs, Tom Goldstein

摘要:

迁移学习,即网络在一个任务上进行训练,并在另一个任务上重新使用,通常用于在数据稀缺或全面训练成本过高时产生神经网络分类器。当目标是生成一个不仅准确而且具有对抗性的模型时,数据稀缺和计算限制变得更加麻烦。我们考虑鲁棒的迁移学习,其中我们不仅将性能而且将鲁棒性从源模型转移到目标域。首先,我们观察到健壮的网络包含健壮的特征提取器。通过在这些特征提取器之上训练分类器,我们产生了继承其父网络鲁棒性的新模型。然后,我们考虑的情况下,微调网络重新训练端到端的目标域。当使用终身学习策略时,此过程在保持源网络的鲁棒性的同时实现高准确性。通过使用这些策略,可以用很少的数据生成准确而强大的模型,并且没有对抗性训练的成本。此外,我们可以提高逆向训练模型的泛化能力,同时保持其鲁棒性。

英文摘要:

Transfer learning, in which a network is trained on one task and re-purposed on another, is often used to produce neural network classifiers when data is scarce or full-scale training is too costly. When the goal is to produce a model that is not only accurate but also adversarially robust, data scarcity and computational limitations become even more cumbersome. We consider robust transfer learning, in which we transfer not only performance but also robustness from a source model to a target domain. We start by observing that robust networks contain robust feature extractors. By training classifiers on top of these feature extractors, we produce new models that inherit the robustness of their parent networks. We then consider the case of fine tuning a network by re-training end-to-end in the target domain. When using lifelong learning strategies, this process preserves the robustness of the source network while achieving high accuracy. By using such strategies, it is possible to produce accurate and robust models with little data, and without the cost of adversarial training. Additionally, we can improve the generalization of adversarially trained models, while maintaining their robustness.

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。

↑