Please use this identifier to cite or link to this item:
http://hdl.handle.net/10397/105517
Title: | Sketch-then-edit generative adversarial network | Authors: | Li, W Xu, L Liang, Z Wang, S Cao, J Ma, C Cui, X |
Issue Date: | 5-Sep-2020 | Source: | Knowledge-based systems, 5 Sept 2020, v. 203, 106102 | Abstract: | Generative Adversarial Network (GAN) has been widely used to generate impressively plausible data. However, it is a non-trivial task to train the original GAN model in practice due to the vanishing gradient problem. This is because the JS divergence could be a constant (i.e., log2) when original data distribution and generated data distribution hold a negligible overlapping area. Under such a scenario, the gradient of generator is 0. Most efforts have been devoted to designing a more proper difference measure while few attentions have been paid to the former aspect of the issue. In this paper, we propose a new method to design a noise distribution having a guaranteed non-negligible overlapping area with raw data distribution. The key idea is to transform the noise from the randomized space into the raw data space. We propose to obtain the transformation as the basis matrix in non-negative matrix factorization because the basis matrix has the underlying features of the raw data. The proposed idea is instantiated as Sketch-then-Edit GAN (SEGAN) where sketches are the noises after transformation and are adopted as the name since they contains basic features of the raw data. Moreover, a new generator for editing the sketches into realistic-like data is designed. We mathematically prove that SEGAN solves the gradient vanishing problem, and conduct extensive experiments on the MNIST, CIFAR10, SVHN and Celeba datasets to demonstrate the effectiveness of SEGAN. |
Keywords: | Generative adversarial network Non-negative matrix factorization Vanishing gradient problem |
Publisher: | Elsevier BV | Journal: | Knowledge-based systems | ISSN: | 0950-7051 | EISSN: | 1872-7409 | DOI: | 10.1016/j.knosys.2020.106102 | Rights: | ©2020 Elsevier B.V. All rights reserved. ©2020. This manuscript version is made available under the CC-BY-NC-ND 4.0 license https://creativecommons.org/licenses/by-nc-nd/4.0/ The following publication Li, W., Xu, L., Liang, Z., Wang, S., Cao, J., Ma, C., & Cui, X. (2020). Sketch-then-edit generative adversarial network. Knowledge-Based Systems, 203, 106102 is available at https://doi.org/10.1016/j.knosys.2020.106102. |
Appears in Collections: | Journal/Magazine Article |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
Xu_Sketch-Then-Edit_Generative_Adversarial.pdf | Pre-Published version | 2.29 MB | Adobe PDF | View/Open |
Page views
9
Citations as of Jul 7, 2024
Downloads
3
Citations as of Jul 7, 2024
SCOPUSTM
Citations
16
Citations as of Jul 4, 2024
WEB OF SCIENCETM
Citations
15
Citations as of Jul 4, 2024
Google ScholarTM
Check
Altmetric
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.