ImageNet pre-training has enabled state-of-the-art results on many tasks. In spite of its recognized contribution to generalization, we observed in this study that ImageNet pre-training also transfers adversarial non-robustness from pre-trained model into fine-tuned model in the downstream classification tasks. We first conducted experiments on various datasets and network backbones to uncover the adversarial non-robustness in fine-tuned model. Further analysis was conducted on examining the learned knowledge of fine-tuned model and standard model, and revealed that the reason leading to the non-robustness is the non-robust features transferred from ImageNet pre-trained model. Finally, we analyzed the preference for feature learning of the pre-trained model, explored the factors influencing robustness, and introduced a simple robust ImageNet pre-training solution. Our code is available at \url{https://github.com/jiamingzhang94/ImageNet-Pretraining-transfers-non-robustness}.
翻译:培训前的图像网络使许多任务得以取得最先进的成果。我们在本研究报告中看到,尽管已经认识到它有助于普及,但我们在研究中注意到,培训前的图像网络还将对抗性非野蛮性从预先培训的模型转变为下游分类任务的微调模型。我们首先在各种数据集和网络主干线上进行了实验,以在微调模型中发现对抗性非野蛮性。还进一步分析了对精细调整模型和标准模型所学知识的研究,并发现导致非野蛮性的原因是从图像网络预培训模型中转移的非野蛮性特征。最后,我们分析了预先培训模型对特征学习的偏好,探讨了影响稳健性的因素,并引入了简单的稳健的图像网络前培训解决方案。我们的代码可在以下网站查阅:url{https://github.com/jiamingzhang94/IMageNet-praining-transfors-nrobastness}。