{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,7,28]],"date-time":"2025-07-28T21:49:43Z","timestamp":1753739383216},"publisher-location":"California","reference-count":0,"publisher":"International Joint Conferences on Artificial Intelligence Organization","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2019,8]]},"abstract":"<jats:p>Expectation maximization (EM) algorithm is to\u00a0find maximum likelihood solution for models having latent variables. A typical example is Gaussian\u00a0Mixture Model (GMM) which requires Gaussian\u00a0assumption, however, natural images are highly\u00a0non-Gaussian so that GMM cannot be applied\u00a0to perform image clustering task on pixel space.\u00a0To overcome such limitation, we propose a GAN\u00a0based EM learning framework that can maximize\u00a0the likelihood of images and estimate the latent\u00a0variables. We call this model GAN-EM, which is\u00a0a framework for image clustering, semi-supervised\u00a0classification and dimensionality reduction. In M-step, we design a novel loss function for discriminator of GAN to perform maximum likelihood estimation (MLE) on data with soft class label assignments. Specifically, a conditional generator captures data distribution for K classes, and a discriminator tells whether a sample is real or fake for\u00a0each class. Since our model is unsupervised, the\u00a0class label of real data is regarded as latent variable,\u00a0which is estimated by an additional network (E-net)\u00a0in E-step. The proposed GAN-EM achieves state-of-the-art clustering and semi-supervised classification results on MNIST, SVHN and CelebA, as\u00a0well as comparable quality of generated images to\u00a0other recently developed generative models.<\/jats:p>","DOI":"10.24963\/ijcai.2019\/612","type":"proceedings-article","created":{"date-parts":[[2019,7,28]],"date-time":"2019-07-28T03:46:05Z","timestamp":1564285565000},"page":"4404-4411","source":"Crossref","is-referenced-by-count":10,"title":["GAN-EM: GAN Based EM Learning Framework"],"prefix":"10.24963","author":[{"given":"Wentian","family":"Zhao","sequence":"first","affiliation":[{"name":"University of Rochester"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Shaojie","family":"Wang","sequence":"additional","affiliation":[{"name":"University of Rochester"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Zhihuai","family":"Xie","sequence":"additional","affiliation":[{"name":"Tsinghua University"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Jing","family":"Shi","sequence":"additional","affiliation":[{"name":"University of Rochester"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Chenliang","family":"Xu","sequence":"additional","affiliation":[{"name":"University of Rochester"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"10584","event":{"number":"28","sponsor":["International Joint Conferences on Artificial Intelligence Organization (IJCAI)"],"acronym":"IJCAI-2019","name":"Twenty-Eighth International Joint Conference on Artificial Intelligence {IJCAI-19}","start":{"date-parts":[[2019,8,10]]},"theme":"Artificial Intelligence","location":"Macao, China","end":{"date-parts":[[2019,8,16]]}},"container-title":["Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence"],"original-title":[],"deposited":{"date-parts":[[2019,7,28]],"date-time":"2019-07-28T03:50:28Z","timestamp":1564285828000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.ijcai.org\/proceedings\/2019\/612"}},"subtitle":[],"proceedings-subject":"Artificial Intelligence Research Articles","short-title":[],"issued":{"date-parts":[[2019,8]]},"references-count":0,"URL":"https:\/\/doi.org\/10.24963\/ijcai.2019\/612","relation":{},"subject":[],"published":{"date-parts":[[2019,8]]}}}