{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,29]],"date-time":"2026-09-29T15:39:24Z","timestamp":1790696364381,"version":"4.1.0"},"reference-count":47,"publisher":"Oxford University Press (OUP)","issue":"1","license":[{"start":{"date-parts":[[2023,2,24]],"date-time":"2023-02-24T00:00:00Z","timestamp":1677196800000},"content-version":"vor","delay-in-days":466,"URL":"https:\/\/academic.oup.com\/journals\/pages\/open_access\/funder_policies\/chorus\/standard_publication_model"}],"funder":[{"DOI":"10.13039\/100000001","name":"National Science Foundation","doi-asserted-by":"publisher","id":[{"id":"10.13039\/100000001","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2022,2,1]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>Generative adversarial networks (GANs) have been impactful on many problems and applications but suffer from unstable training. The Wasserstein GAN (WGAN) leverages the Wasserstein distance to avoid the caveats in the minmax two-player training of GANs but has other defects such as mode collapse and lack of metric to detect the convergence. We introduce a novel inferential Wasserstein GAN (iWGAN) model, which is a principled framework to fuse autoencoders and WGANs. The iWGAN model jointly learns an encoder network and a generator network motivated by the iterative primal-dual optimization process. The encoder network maps the observed samples to the latent space and the generator network maps the samples from the latent space to the data space. We establish the generalization error bound of the iWGAN to theoretically justify its performance. We further provide a rigorous probabilistic interpretation of our model under the framework of maximum likelihood estimation. The iWGAN, with a clear stopping criteria, has many advantages over other autoencoder GANs. The empirical experiments show that the iWGAN greatly mitigates the symptom of mode collapse, speeds up the convergence, and is able to provide a measurement of quality check for each individual sample. We illustrate the ability of the iWGAN by obtaining competitive and stable performances for benchmark datasets.<\/jats:p>","DOI":"10.1111\/rssb.12476","type":"journal-article","created":{"date-parts":[[2021,11,16]],"date-time":"2021-11-16T00:29:54Z","timestamp":1637022594000},"page":"83-113","source":"Crossref","is-referenced-by-count":21,"title":["Inferential Wasserstein Generative Adversarial Networks"],"prefix":"10.1093","volume":"84","author":[{"given":"Yao","family":"Chen","sequence":"first","affiliation":[{"name":"Department of Statistics, Purdue University , West Lafayette, Indiana , USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Qingyi","family":"Gao","sequence":"additional","affiliation":[{"name":"Department of Statistics, Purdue University , West Lafayette, Indiana , USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Xiao","family":"Wang","sequence":"additional","affiliation":[{"name":"Department of Statistics, Purdue University , West Lafayette, Indiana , USA"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"286","published-online":{"date-parts":[[2021,11,15]]},"reference":[{"key":"2023022501100625300_","first-page":"214","volume-title":"Wasserstein generative adversarial networks","author":"Arjovsky","year":"2017"},{"key":"2023022501100625300_","first-page":"224","article-title":"Generalization and equilibrium in generative adversarial nets (gans)","volume-title":"Proceedings of the 34th International Conference on Machine Learning-Volume 70","author":"Arora","year":"2017"},{"key":"2023022501100625300_","article-title":"A note on the inception score","volume-title":"arXiv preprint arXiv:1801.01973","author":"Barratt","year":"2018"},{"key":"2023022501100625300_","first-page":"6240","volume-title":"Spectrally-normalized margin bounds for neural networks","author":"Bartlett","year":"2017"},{"key":"2023022501100625300_","volume-title":"Began: boundary equilibrium generative adversarial networks","author":"Berthelot","year":"2017"},{"issue":"518","key":"2023022501100625300_","doi-asserted-by":"crossref","first-page":"859","DOI":"10.1080\/01621459.2017.1285773","article-title":"Variational inference: a review for statisticians","volume":"112","author":"Blei","year":"2017","journal-title":"Journal of the American Statistical Association"},{"key":"2023022501100625300_","volume-title":"Large scale GAN training for high fidelity natural image synthesis","author":"Brock","year":"2019"},{"key":"2023022501100625300_","first-page":"33","volume":"10","author":"Carreira-Perpinan","year":"2005","journal-title":"On contrastive divergence learning"},{"key":"2023022501100625300_","volume-title":"Training generative adversarial networks via primal-dual subgradient methods: a lagrangian perspective on GAN","author":"Chen","year":"2018"},{"key":"2023022501100625300_","volume-title":"Adversarial feature learning","author":"Donahue","year":"2017"},{"key":"2023022501100625300_","volume-title":"Adversarially learned inference","author":"Dumoulin","year":"2017"},{"key":"2023022501100625300_","first-page":"5248","volume-title":"A convex duality framework for gans","author":"Farnia","year":"2018"},{"key":"2023022501100625300_","first-page":"208","volume-title":"Empirical analysis of the divergence of gibbs sampling based learning algorithms for restricted boltzmann machines","author":"Fischer","year":"2010"},{"key":"2023022501100625300_","first-page":"7515","author":"Gao","year":"2020"},{"key":"2023022501100625300_","first-page":"2672","volume-title":"Generative adversarial nets","author":"Goodfellow","year":"2014"},{"issue":"Mar","key":"2023022501100625300_","first-page":"723","article-title":"A kernel two-sample test","volume":"13","author":"Gretton","year":"2012","journal-title":"Journal of Machine Learning Research"},{"key":"2023022501100625300_","volume-title":"Evaluating gans via duality","author":"Grnarova","year":"2018"},{"issue":"2","key":"2023022501100625300_","doi-asserted-by":"crossref","first-page":"339","DOI":"10.1111\/1467-9868.00289","article-title":"Maximum likelihood estimation for spatial models by markov chain monte carlo stochastic approximation","volume":"63","author":"Gu","year":"2001","journal-title":"Journal of the Royal Statistical Society B"},{"key":"2023022501100625300_","first-page":"5767","volume-title":"Improved training of wasserstein gans","author":"Gulrajani","year":"2017"},{"key":"2023022501100625300_","first-page":"1137","volume-title":"Isometric embeddings of riemannian manifolds","author":"G\u00fcnther","year":"1991"},{"key":"2023022501100625300_","first-page":"8670","volume-title":"Divergence triangle for joint training of generator model, energy-based model, and inferential model","author":"Han","year":"2019"},{"key":"2023022501100625300_","first-page":"6626","volume-title":"Gans trained by a two time-scale update rule converge to a local nash equilibrium","author":"Heusel","year":"2017"},{"issue":"8","key":"2023022501100625300_","doi-asserted-by":"crossref","first-page":"1771","DOI":"10.1162\/089976602760128018","article-title":"Training products of experts by minimizing contrastive divergence","volume":"14","author":"Hinton","year":"2002","journal-title":"Neural Computation"},{"issue":"2","key":"2023022501100625300_","doi-asserted-by":"crossref","first-page":"251","DOI":"10.1016\/0893-6080(91)90009-T","article-title":"Approximation capabilities of multilayer feedforward networks","volume":"4","author":"Hornik","year":"1991","journal-title":"Neural Networks"},{"key":"2023022501100625300_","volume-title":"On unifying deep generative models","author":"Hu","year":"2018"},{"key":"2023022501100625300_","volume-title":"On computation and generalization of generative adversarial networks under spectrum control","author":"Jiang","year":"2019"},{"key":"2023022501100625300_","volume-title":"Adam: A method for stochastic optimization","author":"Kingma","year":"2015"},{"key":"2023022501100625300_","volume-title":"Auto-Encoding Variational Bayes","author":"Kingma","year":"2014"},{"key":"2023022501100625300_","first-page":"1558","volume-title":"Autoencoding beyond pixels using a learned similarity metric","author":"Larsen","year":"2016"},{"key":"2023022501100625300_","first-page":"1718","volume-title":"Generative moment matching networks","author":"Li","year":"2015"},{"key":"2023022501100625300_","volume-title":"On tighter generalization bounds for deep neural networks: CNNs, resnets, and beyond","author":"Li","year":"2019"},{"key":"2023022501100625300_","first-page":"2391","volume-title":"Adversarial variational bayes: Unifying variational autoencoders and generative adversarial networks","author":"Mescheder","year":"2017"},{"key":"2023022501100625300_","volume-title":"Foundations of machine learning","author":"Mohri","year":"2018"},{"issue":"1","key":"2023022501100625300_","doi-asserted-by":"crossref","first-page":"20","DOI":"10.2307\/1969989","article-title":"The imbedding problem for riemannian manifolds","volume":"63","author":"Nash","year":"1956","journal-title":"Annals of Mathematics"},{"key":"2023022501100625300_","first-page":"271","volume-title":"f-gan: Training generative neural samplers using variational divergence minimization","author":"Nowozin","year":"2016"},{"issue":"0","key":"2023022501100625300_","first-page":"1","article-title":"Almond: adaptive latent modeling and optimization via neural networks and langevin diffusion","volume":"0","author":"Qiu","year":"2020","journal-title":"Journal of the American Statistical Association"},{"key":"2023022501100625300_","volume-title":"Unbiased contrastive divergence algorithm for training energy-based latent variable models","author":"Qiu","year":"2020"},{"key":"2023022501100625300_","volume-title":"Variational approaches for auto-encoding generative adversarial networks","author":"Rosca","year":"2017"},{"issue":"3","key":"2023022501100625300_","doi-asserted-by":"crossref","first-page":"470","DOI":"10.1214\/aoms\/1177729394","article-title":"Remarks on a multivariate transformation","volume":"23","author":"Rosenblatt","year":"1952","journal-title":"Annals of Mathematical Statistics"},{"key":"2023022501100625300_","first-page":"2234","volume-title":"Improved techniques for training gans","author":"Salimans","year":"2016"},{"key":"2023022501100625300_","volume-title":"Investigating convergence of restricted boltzmann machine learning","author":"Schulz","year":"2010"},{"key":"2023022501100625300_","volume-title":"Wasserstein auto-encoders","author":"Tolstikhin","year":"2018"},{"key":"2023022501100625300_","volume-title":"It takes (only) two: adversarial generator-encoder networks","author":"Ulyanov","year":"2018"},{"key":"2023022501100625300_","volume-title":"Optimal transport: old and new","author":"Villani","year":"2008"},{"key":"2023022501100625300_","volume-title":"Energy-based generative adversarial network","author":"Zhao","year":"2017"},{"key":"2023022501100625300_","volume-title":"The information-autoencoding family: A lagrangian perspective on latent variable generative modeling","author":"Zhao","year":"2018"},{"key":"2023022501100625300_","first-page":"2223","volume-title":"Unpaired image-to-image translation using cycle-consistent adversarial networks","author":"Zhu","year":"2017"}],"container-title":["Journal of the Royal Statistical Society Series B: Statistical Methodology"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/pdf\/10.1111\/rssb.12476","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/full-xml\/10.1111\/rssb.12476","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/academic.oup.com\/jrsssb\/article-pdf\/84\/1\/83\/49323819\/jrsssb_84_1_83.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/jrsssb\/article-pdf\/84\/1\/83\/49323819\/jrsssb_84_1_83.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,11,12]],"date-time":"2023-11-12T13:01:54Z","timestamp":1699794114000},"score":1,"resource":{"primary":{"URL":"https:\/\/academic.oup.com\/jrsssb\/article\/84\/1\/83\/7056079"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2021,11,15]]},"references-count":47,"journal-issue":{"issue":"1","published-online":{"date-parts":[[2021,11,15]]},"published-print":{"date-parts":[[2022,2,1]]}},"URL":"https:\/\/doi.org\/10.1111\/rssb.12476","relation":{},"ISSN":["1369-7412","1467-9868"],"issn-type":[{"value":"1369-7412","type":"print"},{"value":"1467-9868","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2022,2]]},"published":{"date-parts":[[2021,11,15]]}}}