Convolutional neural networks perform well on object recognition because of a number of recent advances: rectified linear units (ReLUs), data augmentation, dropout, and large labelled datasets. Unsupervised data has been proposed as another way to improve performance. Unfortunately, unsupervised pre-training is not used by state-of-the-art methods leading to the following question: Is unsupervised pre-training still useful given recent advances? If so, when? We answer this in three parts: we 1) develop a unsupervised method that incorporates ReLUs and recent unsupervised regularization techniques, 2) analyze the benefits of unsupervised pre-training compared to data augmentation and dropout on CIFAR-10 while varying the ratio of unsupervised to supervised samples, 3) verify our findings on STL-10. We discover unsupervised pre-training, as expected, helps when the ratio of unsupervised to supervised samples is high, and surprisingly, hurts when the ratio is low. We also use unsupervised pre-training with additional color augmentation to achieve near state-of-the-art performance on STL-10.

本文针对卷积神经网络在物体识别方面的表现进行研究，探讨了无监督预训练在近期发展中是否仍具有重要的作用，并通过实验证实了无监督预训练对于样本中无监督数量比例较高时的效益，而当比例较低时不如数据增强、抑制过拟合等技术，此外文中还提到了色彩增强等方式更接近于先前最先进方法的表现。

近期进展下的无监督预训练分析