{"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/pixelgan-autoencoders","title":"PixelGAN Autoencoders","arxiv_id":"1706.00531","date":"2017-06-02","proceeding":"NeurIPS 2017 12","authors":["Alireza Makhzani","Brendan Frey"],"abstract":"In this paper, we describe the \"PixelGAN autoencoder\", a generative\nautoencoder in which the generative path is a convolutional autoregressive\nneural network on pixels (PixelCNN) that is conditioned on a latent code, and\nthe recognition path uses a generative adversarial network (GAN) to impose a\nprior distribution on the latent code. We show that different priors result in\ndifferent decompositions of information between the latent code and the\nautoregressive decoder. For example, by imposing a Gaussian distribution as the\nprior, we can achieve a global vs. local decomposition, or by imposing a\ncategorical distribution as the prior, we can disentangle the style and content\ninformation of images in an unsupervised fashion. We further show how the\nPixelGAN autoencoder with a categorical prior can be directly used in\nsemi-supervised settings and achieve competitive semi-supervised classification\nresults on the MNIST, SVHN and NORB datasets.","url_abs":"http://arxiv.org/abs/1706.00531v1","url_pdf":"http://arxiv.org/pdf/1706.00531v1.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":"pixelgan-autoencoders","repo_url":"https://github.com/anonyme20/nips20","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}}],"tasks":[{"task_slug":"decoder","task_name":"Decoder"},{"task_slug":null,"task_name":"Generative Adversarial Network"},{"task_slug":"unsupervised-image-classification","task_name":"Unsupervised Image Classification"},{"task_slug":"unsupervised-mnist","task_name":"Unsupervised MNIST"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/unsupervised-image-classification-on-mnist","task":"Unsupervised Image Classification","dataset":"MNIST","model":"PixelGAN Autoencoders","rank_in_archive_order":10,"of":10,"metrics":{"Accuracy":"94.73"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/1706.00531","atlas_url":"https://app.syntology.ai/?focus=1706.00531","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}