The study of cross-domain mapping without supervision has recently attracted much attention. Much of the recent progress was enabled by the use of adversarial training as well as cycle constraints.

Making analogies is a key component in human intelligence.

The manuscript considers the problem of unsupervised mapping across domains by use of variational auto-encoders (VAE) instead of the current adversarial training (GAN) approach. The approach extends a recently proposed non-adversarial mapping (NAM) framework extracting a mapping from a pretrained generative model to a target domain with the present addition of a probabilistic encoder which maps the target image y to the latent domain z_y, i.e. based on the optimization given in equation 2 of the manuscript. Quality: The issue of mapping between domains is interesting and the framework presented seems new and useful. It however also appears to be a straightforward extension to NAM but providing computational merits and improved performance in terms of the quality of mapping which could warrant publication. The experimentation is somewhat limited and the presentation of the paper could be improved by proofreading the manuscript.

Several methods have recently been proposed for the Single Image Super-Resolution (SISR) problem. The current methods assume that a single low-resolution image can only yield a single high-resolution image. In addition, all of these methods use low-resolution images that were artificially generated through simple bilinear down-sampling. We argue that, first and foremost, the problem of SISR is an one-to-many mapping problem between the low-resolution and all possible candidate high-resolution images and we address the challenging task of learning how to realistically degrade and down-sample high-resolution images. To circumvent this problem, we propose SR-NAM which utilizes the Non-Adversarial Mapping (NAM) technique. Furthermore, we propose a degradation model that learns how to transform high-resolution images to low-resolution images that resemble realistically taken low-resolution photos. Finally, some qualitative results for the proposed method along with the weaknesses of SR-NAM are included.

The study of cross-domain mapping without supervision has recently attracted much attention. Much of the recent progress was enabled by the use of adversarial training as well as cycle constraints. In a recent paper, it was shown that cross-domain mapping is possible without the use of cycles or GANs. Although promising, this approach suffers from several drawbacks including costly inference and an optimization variable for every training example preventing the method from using large training sets. We present an alternative approach which is able to achieve non-adversarial mapping using a novel form of Variational Auto-Encoder.

The study of cross-domain mapping without supervision has recently attracted much attention. Much of the recent progress was enabled by the use of adversarial training as well as cycle constraints. The practical difficulty of adversarial training motivates research into non-adversarial methods. In a recent paper, it was shown that cross-domain mapping is possible without the use of cycles or GANs. Although promising, this approach suffers from several drawbacks including costly inference and an optimization variable for every training example preventing the method from using large training sets. We present an alternative approach which is able to achieve non-adversarial mapping using a novel form of Variational Auto-Encoder. Our method is much faster at inference time, is able to leverage large datasets and has a simple interpretation.

The study of cross-domain mapping without supervision has recently attracted much attention. Much of the recent progress was enabled by the use of adversarial training as well as cycle constraints. The practical difficulty of adversarial training motivates research into non-adversarial methods. In a recent paper, it was shown that cross-domain mapping is possible without the use of cycles or GANs. Although promising, this approach suffers from several drawbacks including costly inference and an optimization variable for every training example preventing the method from using large training sets. We present an alternative approach which is able to achieve non-adversarial mapping using a novel form of Variational Auto-Encoder. Our method is much faster at inference time, is able to leverage large datasets and has a simple interpretation.