{"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/escaping-from-collapsing-modes-in-a","title":"Escaping from Collapsing Modes in a Constrained Space","arxiv_id":"1808.07258","date":"2018-08-22","proceeding":"ECCV 2018 9","authors":["Chia-Che Chang","Chieh Hubert Lin","Che-Rung Lee","Da-Cheng Juan","Wei Wei","Hwann-Tzong Chen"],"abstract":"Generative adversarial networks (GANs) often suffer from unpredictable\nmode-collapsing during training. We study the issue of mode collapse of\nBoundary Equilibrium Generative Adversarial Network (BEGAN), which is one of\nthe state-of-the-art generative models. Despite its potential of generating\nhigh-quality images, we find that BEGAN tends to collapse at some modes after a\nperiod of training. We propose a new model, called \\emph{BEGAN with a\nConstrained Space} (BEGAN-CS), which includes a latent-space constraint in the\nloss function. We show that BEGAN-CS can significantly improve training\nstability and suppress mode collapse without either increasing the model\ncomplexity or degrading the image quality. Further, we visualize the\ndistribution of latent vectors to elucidate the effect of latent-space\nconstraint. The experimental results show that our method has additional\nadvantages of being able to train on small datasets and to generate images\nsimilar to a given real image yet with variations of designated attributes\non-the-fly.","url_abs":"http://arxiv.org/abs/1808.07258v1","url_pdf":"http://arxiv.org/pdf/1808.07258v1.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":"escaping-from-collapsing-modes-in-a","repo_url":"https://github.com/chang810249/BEGAN-CS","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":null,"task_name":"Generative Adversarial Network"},{"task_slug":"image-generation","task_name":"Image Generation"}],"methods":[{"method_slug":"convolution","method_name":"Convolution"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/image-generation-on-celeba-64x64","task":"Image Generation","dataset":"CelebA 64x64","model":"BEGAN-CS","rank_in_archive_order":32,"of":39,"metrics":{"FID":"34.136"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1808.07258","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}