https://scholars.lib.ntu.edu.tw/handle/123456789/632498
標題: | A Unified View of cGANs with and without Classifiers | 作者: | Chen S.-A Li C.-L HSUAN-TIEN LIN |
公開日期: | 2021 | 卷: | 33 | 起(迄)頁: | 27566-27579 | 來源出版物: | Advances in Neural Information Processing Systems | 摘要: | Conditional Generative Adversarial Networks (cGANs) are implicit generative models which allow to sample from class-conditional distributions. Existing cGANs are based on a wide range of different discriminator designs and training objectives. One popular design in earlier works is to include a classifier during training with the assumption that good classifiers can help eliminate samples generated with wrong classes. Nevertheless, including classifiers in cGANs often comes with a side effect of only generating easy-to-classify samples. Recently, some representative cGANs avoid the shortcoming and reach state-of-the-art performance without having classifiers. Somehow it remains unanswered whether the classifiers can be resurrected to design better cGANs. In this work, we demonstrate that classifiers can be properly leveraged to improve cGANs. We start by using the decomposition of the joint probability distribution to connect the goals of cGANs and classification as a unified framework. The framework, along with a classic energy model to parameterize distributions, justifies the use of classifiers for cGANs in a principled manner. It explains several popular cGAN variants, such as ACGAN, ProjGAN, and ContraGAN, as special cases with different levels of approximations, which provides a unified view and brings new insights to understanding cGANs. Experimental results demonstrate that the design inspired by the proposed framework outperforms state-of-the-art cGANs on multiple benchmark datasets, especially on the most challenging ImageNet. The code is available at https://github.com/sian-chen/PyTorch-ECGAN. © 2021 Neural information processing systems foundation. All rights reserved. |
URI: | https://www.scopus.com/inward/record.uri?eid=2-s2.0-85131886439&partnerID=40&md5=b56ee757b584ecc9c975a7e7f728b639 https://scholars.lib.ntu.edu.tw/handle/123456789/632498 |
ISSN: | 10495258 | SDG/關鍵字: | Adversarial classifications; Benchmark datasets; Conditional distribution; Energy model; Generative model; Joint probability distributions; Side effect; State of the art; State-of-the-art performance; Unified framework; Generative adversarial networks |
顯示於: | 資訊工程學系 |
在 IR 系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。