{"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/multi-objective-reinforced-evolution-in","title":"Multi-Objective Reinforced Evolution in Mobile Neural Architecture Search","arxiv_id":"1901.01074","date":"2019-01-04","proceeding":null,"authors":["Xiangxiang Chu","Bo Zhang","Ruijun Xu","Hailong Ma"],"abstract":"Fabricating neural models for a wide range of mobile devices demands for a\nspecific design of networks due to highly constrained resources. Both evolution\nalgorithms (EA) and reinforced learning methods (RL) have been dedicated to\nsolve neural architecture search problems. However, these combinations usually\nconcentrate on a single objective such as the error rate of image\nclassification. They also fail to harness the very benefits from both sides. In\nthis paper, we present a new multi-objective oriented algorithm called MoreMNAS\n(Multi-Objective Reinforced Evolution in Mobile Neural Architecture Search) by\nleveraging good virtues from both EA and RL. In particular, we incorporate a\nvariant of multi-objective genetic algorithm NSGA-II, in which the search space\nis composed of various cells so that crossovers and mutations can be performed\nat the cell level. Moreover, reinforced control is mixed with a natural\nmutating process to regulate arbitrary mutation, maintaining a delicate balance\nbetween exploration and exploitation. Therefore, not only does our method\nprevent the searched models from degrading during the evolution process, but it\nalso makes better use of learned knowledge. Our experiments conducted in\nSuper-resolution domain (SR) deliver rivalling models compared to some\nstate-of-the-art methods with fewer FLOPS.","url_abs":"http://arxiv.org/abs/1901.01074v3","url_pdf":"http://arxiv.org/pdf/1901.01074v3.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":"multi-objective-reinforced-evolution-in","repo_url":"https://github.com/moremnas/MoreMNAS","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"tf","reach":null},{"paper_slug":"multi-objective-reinforced-evolution-in","repo_url":"https://github.com/2023-MindSpore-1/ms-code-217/tree/main/sr_ea","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"mindspore","reach":null},{"paper_slug":"multi-objective-reinforced-evolution-in","repo_url":"https://github.com/2023-MindSpore-4/Code11/tree/main/sr_ea","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"mindspore","reach":null},{"paper_slug":"multi-objective-reinforced-evolution-in","repo_url":"https://github.com/2023-MindSpore-4/Code7/tree/main/sr_ea","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"mindspore","reach":null}],"tasks":[{"task_slug":"image-classification","task_name":"Image Classification"},{"task_slug":"architecture-search","task_name":"Neural Architecture Search"},{"task_slug":"super-resolution","task_name":"Super-Resolution"},{"task_slug":"image-classification","task_name":"image-classification"}],"methods":[{"method_slug":"lstm","method_name":"LSTM"},{"method_slug":"sigmoid-activation","method_name":"Sigmoid Activation"},{"method_slug":"softmax","method_name":"Softmax"},{"method_slug":"tanh-activation","method_name":"Tanh Activation"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1901.01074","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}