{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/mmd-gan-towards-deeper-understanding-of","title":"MMD GAN: Towards Deeper Understanding of Moment Matching Network","arxiv_id":"1705.08584","date":"2017-05-24","proceeding":"NeurIPS 2017 12","authors":["Chun-Liang Li","Wei-Cheng Chang","Yu Cheng","Yiming Yang","Barnabás Póczos"],"abstract":"Generative moment matching network (GMMN) is a deep generative model that\ndiffers from Generative Adversarial Network (GAN) by replacing the\ndiscriminator in GAN with a two-sample test based on kernel maximum mean\ndiscrepancy (MMD). Although some theoretical guarantees of MMD have been\nstudied, the empirical performance of GMMN is still not as competitive as that\nof GAN on challenging and large benchmark datasets. The computational\nefficiency of GMMN is also less desirable in comparison with GAN, partially due\nto its requirement for a rather large batch size during the training. In this\npaper, we propose to improve both the model expressiveness of GMMN and its\ncomputational efficiency by introducing adversarial kernel learning techniques,\nas the replacement of a fixed Gaussian kernel in the original GMMN. The new\napproach combines the key ideas in both GMMN and GAN, hence we name it MMD GAN.\nThe new distance measure in MMD GAN is a meaningful loss that enjoys the\nadvantage of weak topology and can be optimized via gradient descent with\nrelatively small batch sizes. In our evaluation on multiple benchmark datasets,\nincluding MNIST, CIFAR- 10, CelebA and LSUN, the performance of MMD-GAN\nsignificantly outperforms GMMN, and is competitive with other representative\nGAN works.","url_abs":"http://arxiv.org/abs/1705.08584v3","url_pdf":"http://arxiv.org/pdf/1705.08584v3.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"mmd-gan-towards-deeper-understanding-of","repo_url":"https://github.com/OctoberChang/MMD-GAN","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}},{"paper_slug":"mmd-gan-towards-deeper-understanding-of","repo_url":"https://github.com/ycjungSubhuman/process-arxiv-pdf","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"unanswered"}},{"paper_slug":"mmd-gan-towards-deeper-understanding-of","repo_url":"https://github.com/Elman295/MMD-GAN","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"none","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":"computational-efficiency","task_name":"Computational Efficiency"},{"task_slug":null,"task_name":"Generative Adversarial Network"}],"methods":[{"method_slug":"convolution","method_name":"Convolution"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1705.08584","atlas_url":"https://app.syntology.ai/?focus=1705.08584","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}