[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-120527-en":3,"doc-seo-120527-105":30,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":4,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":27,"seo_description":14,"update_tm":28,"read_time":29},120527,1374391974564,"Clementine","https://ap-avatar.wpscdn.com/avatar/14000253aa45c000a9e?x-image-process=image/resize,m_fixed,w_180,h_180&k=1779874745381141002",8,"Research & Report","Large Language Models Synergize with Automated Machine Learning - Text-to-ML program synthesis for ML workflows","Program synthesis driven by large language models (LLMs) has gained momentum, yet synthesizing machine learning (ML) programs remains difficult due to long, diverse, and complex workflow code. This paper proposes Text-to-ML, a method that decomposes full ML pipelines into compatible parts generated separately by an LLM using textual task descriptions. A dedicated testing technique enables numerical evaluation and optimization beyond binary correctness, while autoML selects top candidates. Experiments across 12 tasks show improvements in 10 tasks and autoML boosts generated program performance.","arXiv :2405 .03727v 3 [ cs . SE] 9 Sep 2024  \nLarge Language Models Synergize with Automated Machine Learning  \nJinglue Xu  \nUniversity of Tokyo  \nJialong Li  \nTokyo Institute of Technology  \nZhen Liu  \nUniversity of Tokyo  \nNagar Anthel Venkatesh Suryanarayanan  \nUniversity of Tokyo  \nGuoyuan Zhou  \nHosei University Institute of Integrated Science and  \nJia Guo  \nHosei University Institute of Integrated Science and  \nHitoshi Iba  \nUniversity of Tokyo  \nKenji Tei  \n[jingluexu@gmail. com](jingluexu@gmail. com)  \n[lijialong@fuji. waseda.jp](lijialong@fuji. waseda.jp)  \n[liu-zhen@g. ecc. u-tokyo. ac.jp](liu-zhen@g. ecc. u-tokyo. ac.jp)  \n[nav-surya@g. ecc. u-tokyo. ac.jp](nav-surya@g. ecc. u-tokyo. ac.jp)  \n[zhouguoyuan@webmail.hzau. edu. cn](zhouguoyuan@webmail.hzau. edu. cn)  \nTechnology  \n[guojia314@gmail. com](guojia314@gmail. com)  \nTechnology  \n[iba@iba.t. u-tokyo. ac.jp](iba@iba.t. u-tokyo. ac.jp)  \n[tei@c.titech. ac.jp](tei@c.titech. ac.jp)  \nTokyo Institute of Technology  \nReviewed on OpenReview: [https: // openreview. net/ forum? id= RDEaIfOiJM](https: // openreview. net/ forum? id= RDEaIfOiJM)  \nAbstract  \nRecently, program synthesis driven by large language models (LLMs) has become increasingly popular. However, program synthesis for machine learning (ML) tasks still poses significant challenges. This paper explores a novel form of program synthesis, targeting ML programs, by combining LLMs and automated machine learning (autoML) . Specifically, our goal is to fully automate the generation and optimization of the code of the entire ML workflow, from data preparation to modeling and post-processing, utilizing only textual descriptions of the ML tasks. To manage the length and diversity of ML programs, we propose to break each ML program into smaller, manageable parts. Each part is generated separately by the LLM, with careful consideration of their compatibilities. To ensure compatibilities, we design a testing technique for ML programs. Unlike traditional program synthesis, which typically relies on binary evaluations (i.e., correct or incorrect), evaluating ML programs necessitates more than just binary judgments. Our approach automates the numerical evaluation and optimization of these programs, selecting the best candidates through autoML techniques. In experiments across various ML tasks, our method outperforms existing methods in 10 out of 12 tasks for generating ML programs. In addition, autoML significantly improves the performance of the generated ML programs. In experiments, given the textual task description, our method, Text-to-ML, generates the complete and optimized ML program in a fully autonomous process. The implementation of our method is available at [https://github.com/JLX0/llm-automl](https://github.com/JLX0/llm-automl).  \n1 Introduction  \n1.1 Background, challenges, and motivation  \nFigure 1: One-click program synthesis for machine learning. This figure presents an example of our qualitative results based on the dataset Learning. Our method, Text-to-ML, transforms a textual task description into an optimized program, encompassing data preparation, modeling, post processing, and hyperparameters. All modules are seamlessly integrated into a single executable program. The entire process, from the user’s input of a textual description to the creation of an end-to-end, optimized, and executable program, is fully automated. The code examples in the figure are folded and abbreviated for readability.  \nProgram synthesis is the automatic creation of computer programs that satisfy high-level specifications (Manna & Waldinger, 1971) . In recent years, autoregressive large language models (LLMs) (Achiam et al. , 2023; Touvron et al., 2023; Anil et al., 2023) have demonstrated impressive performance in program synthesis tasks (Austin et al., 2021; Li et al., 2022; Shinn et al., 2023) . However, existing LLM-based program synthesis methods primarily focus on traditional coding problems, such as interview questi","cbCaiv4DE3ufl1Od","https://ap.wps.com/l/cbCaiv4DE3ufl1Od","pdf",6119061,1,49,"English","en",105,"# Introduction\n## Background, challenges, and motivation\n# One-click program synthesis for ML\n## Text-to-ML workflow transformation\n# Problem: length, diversity, and testing complexity\n## Length\n## Diversity\n## Testing complexity","[{\"question\":\"What does the proposed method Text-to-ML generate from a user input?\",\"answer\":\"Given a textual task description, Text-to-ML generates a complete, optimized ML program end-to-end, including data preparation, modeling, post-processing, and hyperparameters.\"},{\"question\":\"Why are ML program synthesis tasks harder than traditional coding synthesis?\",\"answer\":\"ML programs require lengthy outputs with multiple task-specific components, vary widely across tasks, and demand evaluation that goes beyond binary correct/incorrect judgments.\"},{\"question\":\"How does the method evaluate and optimize candidate ML programs?\",\"answer\":\"It uses a testing technique tailored for ML programs, enabling automated numerical evaluation and optimization, then selects the best candidates using autoML.\"}]","Large Language Models Synergize with Automated Machine Learning - Text-to-ML program synthesis for ML workflows | PDF",1785730505,123,{"code":4,"msg":31,"data":32},"ok",{"site_id":24,"language":23,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"large-language-models-synergize-with-automated-machine-learning-text-to-ml-program-synthesis-for-ml-workflows","",{"@graph":36,"@context":85},[37,54,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/large-language-models-synergize-with-automated-machine-learning-text-to-ml-program-synthesis-for-ml-workflows/120527/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":62,"encodingFormat":61,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-03",true,{"@type":65,"interactionType":66,"userInteractionCount":4},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What does the proposed method Text-to-ML generate from a user input?","Question",{"text":75,"@type":76},"Given a textual task description, Text-to-ML generates a complete, optimized ML program end-to-end, including data preparation, modeling, post-processing, and hyperparameters.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"Why are ML program synthesis tasks harder than traditional coding synthesis?",{"text":80,"@type":76},"ML programs require lengthy outputs with multiple task-specific components, vary widely across tasks, and demand evaluation that goes beyond binary correct/incorrect judgments.",{"name":82,"@type":73,"acceptedAnswer":83},"How does the method evaluate and optimize candidate ML programs?",{"text":84,"@type":76},"It uses a testing technique tailored for ML programs, enabling automated numerical evaluation and optimization, then selects the best candidates using autoML.","https://schema.org",{"og:url":52,"og:type":87,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":89,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]