{"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/stabilizing-gan-training-with-multiple-random","title":"Stabilizing GAN Training with Multiple Random Projections","arxiv_id":"1705.07831","date":"2017-05-22","proceeding":"ICLR 2018 1","authors":["Behnam Neyshabur","Srinadh Bhojanapalli","Ayan Chakrabarti"],"abstract":"Training generative adversarial networks is unstable in high-dimensions as\nthe true data distribution tends to be concentrated in a small fraction of the\nambient space. The discriminator is then quickly able to classify nearly all\ngenerated samples as fake, leaving the generator without meaningful gradients\nand causing it to deteriorate after a point in training. In this work, we\npropose training a single generator simultaneously against an array of\ndiscriminators, each of which looks at a different random low-dimensional\nprojection of the data. Individual discriminators, now provided with restricted\nviews of the input, are unable to reject generated samples perfectly and\ncontinue to provide meaningful gradients to the generator throughout training.\nMeanwhile, the generator learns to produce samples consistent with the full\ndata distribution to satisfy all discriminators simultaneously. We demonstrate\nthe practical utility of this approach experimentally, and show that it is able\nto produce image samples with higher quality than traditional training with a\nsingle discriminator.","url_abs":"http://arxiv.org/abs/1705.07831v2","url_pdf":"http://arxiv.org/pdf/1705.07831v2.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":"stabilizing-gan-training-with-multiple-random","repo_url":"https://github.com/ayanc/rpgan","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}},{"paper_slug":"stabilizing-gan-training-with-multiple-random","repo_url":"https://github.com/mostafaelaraby/gan-random-projection-pytorch","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"unanswered"}}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1705.07831","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}