{"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/190409035","title":"Evolving Deep Neural Networks by Multi-objective Particle Swarm Optimization for Image Classification","arxiv_id":"1904.09035","date":"2019-03-21","proceeding":null,"authors":["Bin Wang","Yanan sun","Bing Xue","Mengjie Zhang"],"abstract":"In recent years, convolutional neural networks (CNNs) have become deeper in\norder to achieve better classification accuracy in image classification.\nHowever, it is difficult to deploy the state-of-the-art deep CNNs for\nindustrial use due to the difficulty of manually fine-tuning the\nhyperparameters and the trade-off between classification accuracy and\ncomputational cost. This paper proposes a novel multi-objective optimization\nmethod for evolving state-of-the-art deep CNNs in real-life applications, which\nautomatically evolves the non-dominant solutions at the Pareto front. Three\nmajor contributions are made: Firstly, a new encoding strategy is designed to\nencode one of the best state-of-the-art CNNs; With the classification accuracy\nand the number of floating point operations as the two objectives, a\nmulti-objective particle swarm optimization method is developed to evolve the\nnon-dominant solutions; Last but not least, a new infrastructure is designed to\nboost the experiments by concurrently running the experiments on multiple GPUs\nacross multiple machines, and a Python library is developed and released to\nmanage the infrastructure. The experimental results demonstrate that the\nnon-dominant solutions found by the proposed algorithm form a clear Pareto\nfront, and the proposed infrastructure is able to almost linearly reduce the\nrunning time.","url_abs":"http://arxiv.org/abs/1904.09035v2","url_pdf":"http://arxiv.org/pdf/1904.09035v2.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":"190409035","repo_url":"https://github.com/wwwbbb8510/cudam","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":null}],"tasks":[{"task_slug":"classification-1","task_name":"Classification"},{"task_slug":"classification","task_name":"General Classification"},{"task_slug":"image-classification","task_name":"Image Classification"},{"task_slug":"image-classification","task_name":"image-classification"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1904.09035","atlas_url":"https://app.syntology.ai/?focus=1904.09035","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}