MMD GAN swMATH ID: 42580 Software Authors: Chun-Liang Li, Wei-Cheng Chang, Yu Cheng, Yiming Yang, Barnabás Póczos Description: MMD GAN: Towards Deeper Understanding of Moment Matching Network. Generative moment matching network (GMMN) is a deep generative model that differs from Generative Adversarial Network (GAN) by replacing the discriminator in GAN with a two-sample test based on kernel maximum mean discrepancy (MMD). Although some theoretical guarantees of MMD have been studied, the empirical performance of GMMN is still not as competitive as that of GAN on challenging and large benchmark datasets. The computational efficiency of GMMN is also less desirable in comparison with GAN, partially due to its requirement for a rather large batch size during the training. In this paper, we propose to improve both the model expressiveness of GMMN and its computational efficiency by introducing adversarial kernel learning techniques, as the replacement of a fixed Gaussian kernel in the original GMMN. The new approach combines the key ideas in both GMMN and GAN, hence we name it MMD GAN. The new distance measure in MMD GAN is a meaningful loss that enjoys the advantage of weak topology and can be optimized via gradient descent with relatively small batch sizes. In our evaluation on multiple benchmark datasets, including MNIST, CIFAR- 10, CelebA and LSUN, the performance of MMD-GAN significantly outperforms GMMN, and is competitive with other representative GAN works. Homepage: https://arxiv.org/abs/1705.08584 Source Code: https://github.com/OctoberChang/MMD-GAN Dependencies: Python Related Software: Wasserstein GAN; Adam; TensorFlow; StyleGAN; ImageNet; CIFAR; PyTorch; BigGAN; pix2pix; CycleGAN; InfoGAN; AttnGAN; Python; ADCME; DiffSharp; LBFGS-B; L-BFGS; L-BFGS-B; StaticArrays.jl; ZXCalculus.jl Cited in: 12 Publications all top 5 Cited by 52 Authors 1 Anirudh, Rushil 1 Benaim, Sagie 1 Biau, Gérard 1 Candès, Emmanuel J. 1 Darve, Eric 1 Diesendruck, Maurice 1 Feng, Hao 1 Fletcher, Preston Thomas 1 Galanti, Tomer 1 Gong, Lihua 1 Grubinger, Thomas 1 Han, T. Yong-Jin 1 Jastrzębski, Stanisław 1 Ji, Yuan 1 Jiao, Jiantao 1 Kailkhura, Bhavya 1 Kanagawa, Motonobu 1 Knop, Szymon 1 Li, Qunwei 1 Liang, Yingbin 1 Liu, Si-Hang 1 Lughofer, Edwin 1 Marukatat, Sanparith 1 Mazur, Marcin 1 Moser, Bernhard Alois 1 Muandet, Krikamol 1 Müller, Peter 1 Natschläger, Thomas 1 Ni, Yang 1 Podolak, Igor T. 1 Romano, Yaniv 1 Saengkyongam, Sorawit 1 Saminger-Platz, Susanne 1 Sangnier, Maxime 1 Sesia, Matteo 1 Shen, Chaomin 1 Spurek, Przemysław 1 Tabor, Jacek 1 Tanielian, Ugo 1 Tse, David N. C. 1 Varshney, Pramod K. 1 Williamson, Sinead A. 1 Wolf, Lior 1 Xiang, Ling-Zhi 1 Xu, Kailai 1 Zellinger, Werner 1 Zhang, Guixu 1 Zhang, Jize 1 Zhang, Yiying 1 Zhou, Nanrun 1 Zhu, Banghua 1 Zhu, Yitan all top 5 Cited in 9 Serials 4 Journal of Machine Learning Research (JMLR) 1 Computer Methods in Applied Mechanics and Engineering 1 IEEE Transactions on Information Theory 1 Physica A 1 Information Sciences 1 Journal of the American Statistical Association 1 Journal of Computational and Graphical Statistics 1 Journal of the Operations Research Society of China 1 SIAM Journal on Mathematics of Data Science all top 5 Cited in 7 Fields 9 Computer science (68-XX) 5 Statistics (62-XX) 1 Calculus of variations and optimal control; optimization (49-XX) 1 Numerical analysis (65-XX) 1 Statistical mechanics, structure of matter (82-XX) 1 Operations research, mathematical programming (90-XX) 1 Game theory, economics, finance, and other social and behavioral sciences (91-XX) Citations by Year