arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2002.02998cs.LGcs.CRcs.CVstat.ML
收藏

Renofeation:一种提高对抗鲁棒性的简单迁移学习方法

Renofeation: A Simple Transfer Learning Method for Improved Adversarial Robustness

Carnegie Mellon University(卡内基梅隆大学);Microsoft Cloud and AI(微软云与人工智能);The University of Texas at Austin(德克萨斯大学奥斯汀分校)

Ting-Wu Chin, Cha Zhang, Diana Marculescu

摘要:

通过从大规模数据集上的预训练模型进行知识转移进行微调是一种广泛传播的方法,可以有效地在小规模数据集上构建模型。在这项工作中,我们证明了最近通过重新训练最后一个线性层为迁移学习设计的对抗攻击可以成功地欺骗通过端到端微调进行迁移学习训练的模型。这引起了许多工业应用的安全问题。相比之下,使用随机初始化训练而没有传输的模型对此类攻击更具鲁棒性,尽管这些模型通常表现出低得多的准确性。为此,我们提出了噪声特征提取,这是一种新的迁移学习方法,可以从随机初始化训练网络,同时实现与微调竞争的干净数据性能。代码可在https://github.com/cmu-enyac/Renofeation获得。

英文摘要:

Fine-tuning through knowledge transfer from a pre-trained model on a large-scale dataset is a widely spread approach to effectively build models on small-scale datasets. In this work, we show that a recent adversarial attack designed for transfer learning via re-training the last linear layer can successfully deceive models trained with transfer learning via end-to-end fine-tuning. This raises security concerns for many industrial applications. In contrast, models trained with random initialization without transfer are much more robust to such attacks, although these models often exhibit much lower accuracy. To this end, we propose noisy feature distillation, a new transfer learning method that trains a network from random initialization while achieving clean-data performance competitive with fine-tuning. Code available at https://github.com/cmu-enyac/Renofeation.

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。

↑