Refine
Has Fulltext
- yes (1)
Year of publication
- 2023 (1)
Document Type
Language
- English (1)
Is part of the Bibliography
- yes (1)
Learning to Unpermute
(2023)
This paper investigates the effectiveness of deep learning models in matching distributions, using the problem of unpermuting images as a case study. Unpermuting an image refers to the process of reversing the effects of image permutation, which involves randomly shuffling the pixels of an image.The research is divided into two phases. The first phase involves the development of a supervised learning model that trains a neural network on a paired data set of original and permuted images.The second phase focuses on the development of an unsupervised deep learning model that trains the
network on unpaired images. The study demonstrates the ability of deep learning models to match distributions and address the problem of unpermuting images. The findings of this study can provide
insights into the general problem of distribution matching in deep learning .