反向训练的深度网络传输更好：图像分类的示例
Adversarially-Trained Deep Nets Transfer Better: Illustration on Image Classification
摘要：
迁移学习已经成为一种强大的方法，用于将图像识别任务上的预训练深度神经网络适应新的领域。这个过程包括在一个功能丰富的大型源数据集上预先训练一个神经网络，冻结编码基本通用图像属性的早期层，然后微调最后几层，以捕获与目标情况相关的特定信息。当只有有限或弱标记的数据可用于新任务时，这种方法特别有用。在这项工作中，我们证明了对抗训练的模型比非对抗训练的模型迁移得更好，特别是在新领域任务只有有限数据的情况下。此外，我们观察到对抗性训练使学习的表示偏向于保留形状，而不是纹理，这会影响源模型的可移植性。最后，通过影响函数的镜头，我们发现转移的对抗训练模型包含更多的人类可识别的语义信息，这至少部分解释了为什么对抗训练模型转移得更好。
英文摘要：
Transfer learning has emerged as a powerful methodology for adapting pre-trained deep neural networks on image recognition tasks to new domains. This process consists of taking a neural network pre-trained on a large feature-rich source dataset, freezing the early layers that encode essential generic image properties, and then fine-tuning the last few layers in order to capture specific information related to the target situation. This approach is particularly useful when only limited or weakly labeled data are available for the new task. In this work, we demonstrate that adversarially-trained models transfer better than non-adversarially-trained models, especially if only limited data are available for the new domain task. Further, we observe that adversarial training biases the learnt representations to retaining shapes, as opposed to textures, which impacts the transferability of the source models. Finally, through the lens of influence functions, we discover that transferred adversarially-trained models contain more human-identifiable semantic information, which explains -- at least partly -- why adversarially-trained models transfer better.
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。
