{"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/bayesian-gan","title":"Bayesian GAN","arxiv_id":"1705.09558","date":"2017-05-26","proceeding":"NeurIPS 2017","authors":["Yunus Saatchi","Andrew Gordon Wilson"],"abstract":"Generative adversarial networks (GANs) can implicitly learn rich\ndistributions over images, audio, and data which are hard to model with an\nexplicit likelihood. We present a practical Bayesian formulation for\nunsupervised and semi-supervised learning with GANs. Within this framework, we\nuse stochastic gradient Hamiltonian Monte Carlo to marginalize the weights of\nthe generator and discriminator networks. The resulting approach is\nstraightforward and obtains good performance without any standard interventions\nsuch as feature matching, or mini-batch discrimination. By exploring an\nexpressive posterior over the parameters of the generator, the Bayesian GAN\navoids mode-collapse, produces interpretable and diverse candidate samples, and\nprovides state-of-the-art quantitative results for semi-supervised learning on\nbenchmarks including SVHN, CelebA, and CIFAR-10, outperforming DCGAN,\nWasserstein GANs, and DCGAN ensembles.","url_abs":"http://arxiv.org/abs/1705.09558v3","url_pdf":"http://arxiv.org/pdf/1705.09558v3.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":"bayesian-gan","repo_url":"https://github.com/andrewgordonwilson/bayesgan","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":null},{"paper_slug":"bayesian-gan","repo_url":"https://github.com/KeAWang/BayesianGAN4AdversarialAttacks","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null},{"paper_slug":"bayesian-gan","repo_url":"https://github.com/rafa2000/Top-TensorFlow","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"bayesian-gan","repo_url":"https://github.com/ranery/Bayesian-CycleGAN","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[],"methods":[{"method_slug":"batch-normalization","method_name":"Batch Normalization"},{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"dcgan","method_name":"DCGAN"},{"method_slug":"feature-matching","method_name":"GAN Feature Matching"},{"method_slug":"relu","method_name":"ReLU"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1705.09558","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}