{"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/operation-level-early-stopping-for","title":"Operation-Level Early Stopping for Robustifying Differentiable NAS","arxiv_id":null,"date":"2023-09-21","proceeding":"NeurIPS 2023 11","authors":[],"abstract":"Differentiable NAS (DARTS) is a simple and efficient neural architecture search method that has been extensively adopted in various machine learning tasks.\n% \nNevertheless, DARTS still encounters several robustness issues, mainly the domination of skip connections.\n% \nThe resulting architectures are full of parametric-free operations, leading to performance collapse.\n% \nExisting methods suggest that the skip connection has additional advantages in optimization compared to other parametric operations and propose to alleviate the domination of skip connections by eliminating these additional advantages.\n% \nIn this paper, we analyze this issue from a simple and straightforward perspective and propose that the domination of skip connections results from parametric operations overfitting the training data while architecture parameters are trained on the validation data, leading to undesired behaviors.\n% \nBased on this observation, we propose the operation-level early stopping (OLES) method to overcome this issue and robustify DARTS without introducing any computation overhead.\n% \nExtensive experimental results can verify our hypothesis and the effectiveness of OLES.Submission Number: 11229","url_abs":"https://openreview.net/forum?id=yAOwkf4FyL","url_pdf":"https://openreview.net/pdf?id=yAOwkf4FyL","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":"operation-level-early-stopping-for","repo_url":"https://github.com/pasalab/oles","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"pytorch","reach":null}],"tasks":[],"methods":[{"method_slug":"darts","method_name":"DARTS"},{"method_slug":"early-stopping","method_name":"Early Stopping"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}