{"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/evolutionary-generative-adversarial-networks","title":"Evolutionary Generative Adversarial Networks","arxiv_id":"1803.00657","date":"2018-03-01","proceeding":null,"authors":["Chaoyue Wang","Chang Xu","Xin Yao","DaCheng Tao"],"abstract":"Generative adversarial networks (GAN) have been effective for learning\ngenerative models for real-world data. However, existing GANs (GAN and its\nvariants) tend to suffer from training problems such as instability and mode\ncollapse. In this paper, we propose a novel GAN framework called evolutionary\ngenerative adversarial networks (E-GAN) for stable GAN training and improved\ngenerative performance. Unlike existing GANs, which employ a pre-defined\nadversarial objective function alternately training a generator and a\ndiscriminator, we utilize different adversarial training objectives as mutation\noperations and evolve a population of generators to adapt to the environment\n(i.e., the discriminator). We also utilize an evaluation mechanism to measure\nthe quality and diversity of generated samples, such that only well-performing\ngenerator(s) are preserved and used for further training. In this way, E-GAN\novercomes the limitations of an individual adversarial training objective and\nalways preserves the best offspring, contributing to progress in and the\nsuccess of GANs. Experiments on several datasets demonstrate that E-GAN\nachieves convincing generative performance and reduces the training problems\ninherent in existing GANs.","url_abs":"http://arxiv.org/abs/1803.00657v1","url_pdf":"http://arxiv.org/pdf/1803.00657v1.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":"evolutionary-generative-adversarial-networks","repo_url":"https://github.com/E-kitcher/MSc-Project-","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":null},{"paper_slug":"evolutionary-generative-adversarial-networks","repo_url":"https://github.com/WANG-Chaoyue/EvolutionaryGAN","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"evolutionary-generative-adversarial-networks","repo_url":"https://github.com/WANG-Chaoyue/EvolutionaryGAN-pytorch","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"diversity","task_name":"Diversity"}],"methods":[{"method_slug":"convolution","method_name":"Convolution"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1803.00657","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}