{"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/optimizing-deep-neural-networks-with-multiple","title":"Optimizing Deep Neural Networks with Multiple Search Neuroevolution","arxiv_id":"1901.05988","date":"2019-01-17","proceeding":null,"authors":["Ahmed Aly","David Weikersdorfer","Claire Delaunay"],"abstract":"This paper presents an evolutionary metaheuristic called Multiple Search\nNeuroevolution (MSN) to optimize deep neural networks. The algorithm attempts\nto search multiple promising regions in the search space simultaneously,\nmaintaining sufficient distance between them. It is tested by training neural\nnetworks for two tasks, and compared with other optimization algorithms. The\nfirst task is to solve Global Optimization functions with challenging\ntopographies. We found to MSN to outperform classic optimization algorithms\nsuch as Evolution Strategies, reducing the number of optimization steps\nperformed by at least 2X.\n  The second task is to train a convolutional neural network (CNN) on the\npopular MNIST dataset. Using 3.33% of the training set, MSN reaches a\nvalidation accuracy of 90%. Stochastic Gradient Descent (SGD) was able to match\nthe same accuracy figure, while taking 7X less optimization steps. Despite\nlagging, the fact that the MSN metaheurisitc trains a 4.7M-parameter CNN\nsuggests promise for future development. This is by far the largest network\never evolved using a pool of only 50 samples.","url_abs":"http://arxiv.org/abs/1901.05988v1","url_pdf":"http://arxiv.org/pdf/1901.05988v1.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":"optimizing-deep-neural-networks-with-multiple","repo_url":"https://github.com/AroMorin/DNNOP","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"global-optimization","task_name":"global-optimization"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}