{"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/first-order-generative-adversarial-networks","title":"First Order Generative Adversarial Networks","arxiv_id":"1802.04591","date":"2018-02-13","proceeding":"ICML 2018 7","authors":["Calvin Seward","Thomas Unterthiner","Urs Bergmann","Nikolay Jetchev","Sepp Hochreiter"],"abstract":"GANs excel at learning high dimensional distributions, but they can update\ngenerator parameters in directions that do not correspond to the steepest\ndescent direction of the objective. Prominent examples of problematic update\ndirections include those used in both Goodfellow's original GAN and the\nWGAN-GP. To formally describe an optimal update direction, we introduce a\ntheoretical framework which allows the derivation of requirements on both the\ndivergence and corresponding method for determining an update direction, with\nthese requirements guaranteeing unbiased mini-batch updates in the direction of\nsteepest descent. We propose a novel divergence which approximates the\nWasserstein distance while regularizing the critic's first order information.\nTogether with an accompanying update direction, this divergence fulfills the\nrequirements for unbiased steepest descent updates. We verify our method, the\nFirst Order GAN, with image generation on CelebA, LSUN and CIFAR-10 and set a\nnew state of the art on the One Billion Word language generation task. Code to\nreproduce experiments is available.","url_abs":"http://arxiv.org/abs/1802.04591v2","url_pdf":"http://arxiv.org/pdf/1802.04591v2.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":"first-order-generative-adversarial-networks","repo_url":"https://github.com/zalandoresearch/first_order_gan","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":"image-generation","task_name":"Image Generation"},{"task_slug":"text-generation","task_name":"Text Generation"}],"methods":[{"method_slug":"convolution","method_name":"Convolution"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/image-generation-on-cifar-10","task":"Image Generation","dataset":"CIFAR-10","model":"FOGAN","rank_in_archive_order":66,"of":78,"metrics":{"FID":"27.4"},"uses_additional_data":false},{"leaderboard":"/sota/image-generation-on-lsun-bedroom-64-x-64","task":"Image Generation","dataset":"LSUN Bedroom 64 x 64","model":"FOGAN","rank_in_archive_order":3,"of":3,"metrics":{"FID":"11.4"},"uses_additional_data":false}],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}