CRITICAL ANALYSIS OF SELF-SUPERVISION
This paper evaluated the performance of the state-of-the-art self-supervised methods on learning weights of convolutional neural networks (CNNs) and on a per-layer basis. They were motivated by the fact that low-level features in the first layers of networks may not require the high-level semantic information captured by manual labels. This paper also aims to figure out whether current self-supervision techniques can learn deep features from only one image.
The main goal of self-supervised learning is to take advantage of a vast amount of unlabeled data to train CNNs and find a generalized image representation. In self-supervised learning, unlabeled data generate ground truth labels per se by pretext tasks such as the Jigsaw puzzle task, and the rotation estimation. For example, in the rotation task, we have a picture of a bird without the label "bird". We rotate the bird image by 90 degrees clockwise and the CNN is trained in a way that to find the rotation axis, as can be seen in the figure below.
In recent literature, several papers addressed self-supervised learning methods and learning from a single sample.
A BiGAN , or Bidirectional GAN, is simply a generative adversarial network plus an encoder. The generator maps latent samples to generated data and the encoder performs as the opposite of the generator. After training BiGAN, the encoder has learned to generate a rich image representation. In RotNet method , images are rotated and the CNN learns to figure out the direction. DeepCluster  alternates k-means clustering to learn stable feature representations under several image transformations.
Method & Experiment
In this paper, BiGAN, RotNet, and DeepCluster are employed for training AlexNet in a self-supervised manner. To evaluate the impact of the size of the training set, they have compared the results of a million images in the ImageNet dataset with a million augmented images generated from only one single image. Various data augmentation methods including cropping, rotation, scaling, contrast changes, and adding noise, have been used to generate the mentioned artificial dataset from one image. Augmentation can be seen as imposing a prior on how we expect the manifold of natural images to look like. When training with very few images, these priors become more important since the model cannot extract them directly from data.
With the intention of measuring the quality of deep features on a per-layer basis, a linear classifier is trained on top of each convolutional layer of AlexNet. Linear classifier probes are commonly used to monitor the features at every layer of a CNN and are trained entirely independently of the CNN itself . Note that the main purpose of CNNs is to reach a linearly discriminable representation for images. Accordingly, the linear probing technique aims to evaluate the training of each layer of a CNN and inspect how much information each of the layers learned. The same experiment has been done using the CIFAR10/100 dataset.
Figure 2 shows how well representations at each level are linearly separable. The result table indicated the classification accuracy of the linear classifier trained on the top of each convolutional layer. According to the results, training the CNN with self-supervision methods can match the performance of fully supervised learning in the first two convolutional layers. It must be pointed out that only one single image with massive augmentation is utilized in this experiment.
The source code for the paper can be found here: https://github.com/yukimasano/linear-probes
In this paper, the authors conduct interesting experiments to show that the first few layers of CNNs contain only limited information for analyzing natural images. They saw this by examining the weights of the early layers in cases where they only trained using only a single image with much data augmentation. Specifically, sufficient data augmentation was enough to make up for a lack of data in early CNN layers. However, this technique was not able to elicit proper learning in deeper CNN layers. In fact, even millions of images were not enough to elicit proper learning without supervision. Accordingly, current unsupervised learning is only about augmentation, and we probably do not use the capacity of a million images, yet.
This is a well-written paper. However, as the main contribution of the paper is experimental, I expected a more in-depth analysis. For example, it is interesting to see how these results change if we change AlexNet with a more powerful CNN like EfficientNet?
 Y. Asano, C. Rupprecht, and A. Vedaldi, “A critical analysis of self-supervision, or what we can learn from a single image,” in International Conference on Learning Representations, 2019
 J. Donahue, P. Kr ̈ahenb ̈uhl, and T. Darrell, “Adversarial feature learning,”arXiv preprint arXiv:1605.09782, 2016.
 S. Gidaris, P. Singh, and N. Komodakis, “Unsupervised representation learning by predicting image rotations,”arXiv preprintarXiv:1803.07728, 2018
 M. Caron, P. Bojanowski, A. Joulin, and M. Douze, “Deep clustering for unsupervised learning of visual features,” in Proceedings of the European Conference on Computer Vision (ECCV), 2018, pp. 132–149
 G. Alain and Y. Bengio, “Understanding intermediate layers using linear classifier probes,”arXiv preprint arXiv:1610.01644, 2016.
 Mehdi Noroozi and Paolo Favaro. Unsupervised learning of visual representations by solving jigsaw puzzles. In ECCV, 2016.