{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,2,20]],"date-time":"2026-02-20T07:07:30Z","timestamp":1771571250152,"version":"3.50.1"},"reference-count":25,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,2,20]],"date-time":"2026-02-20T00:00:00Z","timestamp":1771545600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Artif. Intell."],"abstract":"<jats:p>\n                    The Information Maximizing Generative Adversarial Network (InfoGAN) can be formulated as a minimax problem involving a generator and a discriminator, augmented by a mutual information regularization term. Despite strong empirical performance, rigorous generalization guarantees for InfoGAN-type objectives remain limited, particularly when additional structural components are introduced. In this paper, we study an InfoGAN-inspired adversarial framework obtained by removing the latent code component and introducing an explicit regularization term on the generator, yielding an analytically tractable generator-regularized adversarial objective. We establish generalization error bounds by analyzing the gap between empirical and population objective functions using Rademacher complexity arguments for the discriminator, the generator, and their composition. The resulting bounds reveal explicit\n                    <jats:italic>n<\/jats:italic>\n                    <jats:sup>\u22121\/2<\/jats:sup>\n                    and\n                    <jats:italic>m<\/jats:italic>\n                    <jats:sup>\u22121\/2<\/jats:sup>\n                    decay rates with respect to the discriminator and generator sample sizes and clarify the role of the generator regularization parameter. The theory is further specialized to two-layer neural networks with Lipschitz continuous and non-decreasing activation functions, where explicit entropy-based complexity bounds are derived. Experiments on the CIFAR-10 dataset validate the predicted scaling behavior and demonstrate that the generalization gap decreases systematically as sample size increases, highlighting the stabilizing effect of generator regularization. Overall, this work provides one of the first rigorous generalization analyses for an InfoGAN-inspired adversarial objective with explicit generator regularization.\n                  <\/jats:p>","DOI":"10.3389\/frai.2026.1731256","type":"journal-article","created":{"date-parts":[[2026,2,20]],"date-time":"2026-02-20T06:40:28Z","timestamp":1771569628000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Generalization bounds for a generator-regularized InfoGAN-inspired adversarial objective"],"prefix":"10.3389","volume":"9","author":[{"given":"Mahmud","family":"Hasan","sequence":"first","affiliation":[{"name":"Department of Biostatistics, Virginia Commonwealth University","place":["Richmond, VA, United States"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Mathias Nthiani","family":"Muia","sequence":"additional","affiliation":[{"name":"Department of Mathematics and Statistics, University of South Alabama","place":["Mobile, AL, United States"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Md Mahmudul","family":"Islam","sequence":"additional","affiliation":[{"name":"Department of Mathematics, The University of Alabama at Birmingham","place":["Birmingham, AL, United States"]}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"1965","published-online":{"date-parts":[[2026,2,20]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","DOI":"10.1017\/CBO9780511624216","author":"Anthony","year":"1999","journal-title":"Learning in Neural Networks: Theoretical Foundations"},{"key":"B2","article-title":"\u201cDouble InfoGAN for contrastive analysis,\u201d","volume-title":"Proceedings of the 27th International Conference on Artificial Intelligence and Statistics (AISTATS)","author":"Carton","year":"2024"},{"key":"B3","article-title":"\u201cinfoGAN: interpretable representation learning by information maximizing generative adversarial nets,\u201d","volume-title":"Neural Information Processing Systems (NIPS)","author":"Chen","year":"2016"},{"key":"B4","author":"Dudley","year":"2018","journal-title":"Real Analysis and Probability"},{"key":"B5","first-page":"2672","article-title":"\u201cGenerative adversarial nets,\u201d","volume-title":"Advances in Neural Information Processing Systems (NIPS)","author":"Goodfellow","year":"2014"},{"key":"B6","doi-asserted-by":"publisher","first-page":"3313","DOI":"10.1109\/TKDE.2021.3130191","article-title":"A review on generative adversarial networks: algorithms, theory, and applications","volume":"35","author":"Gui","year":"2023","journal-title":"IEEE Trans. Knowl. Data Eng"},{"key":"B7","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2310.00443","article-title":"Generalization error property of infoGAN for two-layer neural network","author":"Hasan","year":"2025","journal-title":"arXiv [preprint]"},{"key":"B8","first-page":"1","article-title":"An error analysis of generative adversarial networks for learning distributions","volume":"23","author":"Huang","year":"2022","journal-title":"J. Mach. Learn. Res"},{"key":"B9","doi-asserted-by":"publisher","first-page":"14","DOI":"10.1007\/s10994-025-06953-4","article-title":"VE-cGAN: improved generalization analysis of conditional GANs. machine learning","volume":"115","author":"Jang","year":"2026","journal-title":"Mach. Learn"},{"key":"B10","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv:2510.20165","article-title":"IB-GAN: disentangled representation learning with information bottleneck generative adversarial networks","author":"Jeon","year":"2025","journal-title":"arXiv [preprint]"},{"key":"B11","doi-asserted-by":"publisher","first-page":"3114","DOI":"10.1109\/TIT.2021.3053234","article-title":"Understanding estimation and generalization error of generative adversarial networks","volume":"67","author":"Ji","year":"2021","journal-title":"IEEE Trans. Inf. Theory"},{"key":"B12","first-page":"8733","article-title":"\u201cLearning plannable representations with causal info gan,\u201d","volume-title":"Advances in Neural Information Processing Systems (NIPS)","author":"Kurutach","year":"2018"},{"key":"B13","doi-asserted-by":"publisher","first-page":"1","DOI":"10.2139\/ssrn.3714011","article-title":"How well generative adversarial networks learn distributions","volume":"22","author":"Liang","year":"2021","journal-title":"J. Mach. Learn. Res"},{"key":"B14","article-title":"Conditional generative adversarial nets","author":"Mirza","year":"2014","journal-title":"arXiv [preprint]"},{"key":"B15","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2404.00521","article-title":"CHAIN: enhancing generalization in data-efficient GANs via lipsCHitz continuity constrAIned normalization","author":"Ni","year":"2024","journal-title":"arXiv [preprint]"},{"key":"B16","doi-asserted-by":"publisher","first-page":"866","DOI":"10.1016\/j.amc.2018.07.030","article-title":"The epidemic spreading on the multi-relationships network","volume":"339","author":"Nian","year":"2018","journal-title":"Appl. Math. Comput"},{"key":"B17","first-page":"271","article-title":"\u201cF-GAN: training generative neural samplers using variational divergence minimization,\u201d","volume-title":"Advances in Neural Information Processing Systems (NIPS)","author":"Nowozin","year":"2016"},{"key":"B18","volume-title":"Neural Network Theory","author":"Petersen","year":"2022"},{"key":"B19","article-title":"\u201cGenerative adversarial text to image synthesis,\u201d","volume-title":"Proceedings of The 33rd International Conference on Machine Learning (ICML)","author":"Reed","year":"2016"},{"key":"B20","first-page":"1024","article-title":"\u201cNonparametric density estimation under adversarial losses,\u201d","volume-title":"Advances in Neural Information Processing Systems","author":"Singh","year":"2018"},{"key":"B21","doi-asserted-by":"publisher","first-page":"134","DOI":"10.1016\/j.amc.2018.12.045","article-title":"The impact of awareness diffusion on SIR-like epidemics in multiplex networks","volume":"349","author":"Wang","year":"2019","journal-title":"Appl. Math. Comput"},{"key":"B22","article-title":"Logan: latent optimization for generative adversarial networks","author":"Wu","year":"2019","journal-title":"arXiv [preprint]"},{"key":"B23","doi-asserted-by":"publisher","first-page":"101552","DOI":"10.1016\/j.media.2019.101552","article-title":"Generative adversarial network in medical imaging: a review","volume":"58","author":"Yi","year":"2019","journal-title":"Med. Image Anal"},{"key":"B24","article-title":"\u201cOn the discrimination - generalization trade-off in GANs,\u201d","volume-title":"Proceedings International Conference on Learning Representations (ICLR)","author":"Zhang","year":"2018"},{"key":"B25","doi-asserted-by":"crossref","DOI":"10.1109\/ICCV.2017.244","article-title":"\u201cUnpaired image-to-image translation using cycle-consistent adversarial networks,\u201d","volume-title":"IEEE International Conference on Computer Vision","author":"Zhu","year":"2017"}],"container-title":["Frontiers in Artificial Intelligence"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frai.2026.1731256\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,2,20]],"date-time":"2026-02-20T06:40:32Z","timestamp":1771569632000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frai.2026.1731256\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,2,20]]},"references-count":25,"alternative-id":["10.3389\/frai.2026.1731256"],"URL":"https:\/\/doi.org\/10.3389\/frai.2026.1731256","relation":{},"ISSN":["2624-8212"],"issn-type":[{"value":"2624-8212","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,2,20]]},"article-number":"1731256"}}