[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-1-en-105":3,"doc-seo-245319-105":53,"doc-detail-245319-en":127},{"code":4,"msg":5,"data":6},0,"success",[7,14,19,24,29,34,39,44,49],{"id":8,"doc_module":9,"doc_module_name":10,"category_name":11,"show_sort_weight":12,"slug":13},11,1,"Template","Presentations",90,"presentations",{"id":15,"doc_module":9,"doc_module_name":10,"category_name":16,"show_sort_weight":17,"slug":18},12,"Resumes",80,"resumes",{"id":20,"doc_module":9,"doc_module_name":10,"category_name":21,"show_sort_weight":22,"slug":23},14,"Invoices",70,"invoices",{"id":25,"doc_module":9,"doc_module_name":10,"category_name":26,"show_sort_weight":27,"slug":28},15,"Posters",60,"posters",{"id":30,"doc_module":9,"doc_module_name":10,"category_name":31,"show_sort_weight":32,"slug":33},16,"Social Media",50,"social-media",{"id":35,"doc_module":9,"doc_module_name":10,"category_name":36,"show_sort_weight":37,"slug":38},17,"Forms",40,"forms",{"id":40,"doc_module":9,"doc_module_name":10,"category_name":41,"show_sort_weight":42,"slug":43},18,"Letters",30,"letters",{"id":45,"doc_module":9,"doc_module_name":10,"category_name":46,"show_sort_weight":47,"slug":48},21,"Paper Templates",5,"papers-templates",{"id":50,"doc_module":9,"doc_module_name":10,"category_name":51,"show_sort_weight":4,"slug":52},158,"General","general-158",{"code":4,"msg":54,"data":55},"ok",{"site_id":56,"language":57,"slug":58,"title":59,"keywords":60,"description":61,"schema_data":62,"social_meta":120,"head_meta":122,"extra_data":124,"updated_unix":126},105,"en","scaling-manual-grounded-appliance-manipulation-with-data-synthesis-and-unified-planning","Scaling Manual-Grounded Appliance Manipulation with Data Synthesis and Unified Planning","","Operating household appliances requires long-horizon planning that is state-dependent and robust to disturbances, yet existing large models fall short due to the lack of a sufficiently diverse, task-oriented dataset. To bridge this gap, MAGE proposes a scalable data synthesis pipeline with a Hierarchical Appliance Graph to generate grounding, long-horizon planning, and closed-loop recovery data from manuals. UseAppliance spans 22 categories with 89K+ part annotations and 53K+ tasks. Built on it, AppliancePlan achieves over 10× improvements and strong sim-to-real transfer on real robots.",{"@graph":63,"@context":119},[64,80,102],{"@type":65,"itemListElement":66},"BreadcrumbList",[67,71,74,77],{"item":68,"name":69,"@type":70,"position":9},"https://docshare.wps.com","Home","ListItem",{"item":72,"name":10,"@type":70,"position":73},"https://docshare.wps.com/template/",2,{"item":75,"name":11,"@type":70,"position":76},"https://docshare.wps.com/template/presentations/",3,{"item":78,"name":59,"@type":70,"position":79},"https://docshare.wps.com/template/scaling-manual-grounded-appliance-manipulation-with-data-synthesis-and-unified-planning/245319/",4,{"url":78,"name":59,"@type":81,"image":82,"author":87,"headline":59,"publisher":90,"fileFormat":93,"inLanguage":57,"description":61,"dateModified":94,"datePublished":95,"encodingFormat":93,"isAccessibleForFree":96,"interactionStatistic":97},"DigitalDocument",{"url":83,"@type":84,"width":85,"height":86},"https://docshare.wps.com/thumbnails/scaling-manual-grounded-appliance-manipulation-with-data-synthesis-and-unified-planning/245319.png","ImageObject",442,249,{"name":88,"@type":89},"Paura","Person",{"url":68,"name":91,"@type":92},"DocShare","Organization","application/pdf","2026-10-19","2026-09-12",true,{"@type":98,"interactionType":99,"userInteractionCount":101},"InteractionCounter",{"@type":100},"ViewAction",8,{"@type":103,"mainEntity":104},"FAQPage",[105,111,115],{"name":106,"@type":107,"acceptedAnswer":108},"Why do existing large models struggle with appliance manipulation planning?","Question",{"text":109,"@type":110},"They require state-dependent, long-horizon procedures that remain robust under disturbances, but there is no sufficiently diverse, task-oriented dataset tailored to manual-grounded manipulation.","Answer",{"name":112,"@type":107,"acceptedAnswer":113},"What is MAGE and how does it generate training data?",{"text":114,"@type":110},"MAGE is a scalable data synthesis pipeline that uses a Hierarchical Appliance Graph to represent manuals and generates grounding, long-horizon planning, and closed-loop recovery data via graph traversal, disturbance injection, and human verification.",{"name":116,"@type":107,"acceptedAnswer":117},"What are UseAppliance and AppliancePlan in this work?",{"text":118,"@type":110},"UseAppliance is a large-scale dataset built from appliance manuals across 22 categories, containing part annotations, manipulation tasks, and closed-loop steps. AppliancePlan is an end-to-end model for manual-grounded appliance manipulation planning trained on UseAppliance.","https://schema.org",{"og:url":78,"og:type":121,"og:title":59,"og:site_name":91,"og:description":61},"article",{"robots":123,"canonical":78},"index,follow",{"doc_id":125,"site_id":56},245319,1791720372,{"code":4,"msg":5,"data":128},{"doc_id":125,"user_id":129,"nickname":88,"user_avatar":130,"doc_module":9,"category_id":8,"category_name":11,"doc_title":59,"doc_description":61,"doc_content":131,"file_id":132,"file_url":133,"file_type":134,"file_size":135,"view_count":101,"is_deleted":4,"is_public":9,"is_downloadable":9,"audit_status":9,"page_count":136,"language":137,"language_code":57,"site_id":56,"html_lang":57,"table_of_contents":138,"faqs":139,"seo_title":140,"seo_description":61,"update_tm":141,"read_time":79},13056712833777,"https://ap-avatar.wpscdn.com/davatar_29158cc5080c5b710cf443261637dec0","Scaling Manual-Grounded Appliance Manipulation with Data  \nSynthesis and Unified Planning  \nYuxing Long∗ Center on Frontiers of Computing Studies, School of Computer Science, Peking University  \nLei Kang∗ Center on Frontiers of Computing Studies, School of Computer Science, Peking University  \nZiyan Yu  \nCenter on Frontiers of Computing Studies, School of Computer Science, Peking University  \nYuzheng Gao  \nCenter on Frontiers of Computing Studies, School of Computer Science, Peking University  \nBin Cheng  \nBeijing University of Aeronautics and Astronautics  \nJiyao Zhang  \nCenter on Frontiers of Computing Studies, School of Computer Science, Peking University  \narXiv :2608 . 15863v1 [ cs .RO] 16 Aug 2026  \nXiaoqi Li  \nCenter on Frontiers of Computing Studies, School of Computer Science, Peking University  \nHaolin Yang  \nCenter on Frontiers of Computing Studies, School of Computer Science, Peking University  \nDongjiang Li  \nJingdong Technology Information Technology Co., Ltd  \nHui Shen  \nJingdong Technology Information Technology Co., Ltd  \nHao Dong† Center on Frontiers of Computing Studies, School of Computer Science, Peking University [hao.dong@pku.edu.cn](hao.dong@pku.edu.cn)  \nAbstract  \nOperating household appliances requires long-horizon planning that is state-dependent and robust to disturbances, yet existing large models fall short, as no sufficiently diverse, task-oriented dataset exists to support such planning. To bridge this gap, we propose MAGE, a scalable data synthesis pipeline that introduces a novel Hierarchical Appliance Graph (HAG) to automatically generate part grounding, long-horizon planning, and closed-loop recovery data from appliance manuals. With MAGE, we build UseAppliance, the first large-scale dataset for manual-grounded appliance manipulation planning, spanning 22 appliance categories with 89K+ part annotations, 53K+ manipulation tasks, and 33K+ closed-loop adjustment steps. Built on UseAppliance, we develop AppliancePlan, an end-to-end model for manual-grounded appliance manipulation planning. On RealAppliance-Bench, AppliancePlan with only 7B parameters achieves over 10× the best baseline on open-loop planning and consistently outperforms state-of-the-art models across all tasks. Real-robot experiments on six household appliances further confirm effective sim-to-real transfer, marking an important step toward general-purpose household robotics.  \n∗ These authors contributed equally to this work.†Corresponding author.  \nThis work is licensed under a Creative Commons Attribution 4 .0 International License. MM’26, Rio de Janeiro, Brazil  \n© 2026 Copyright held by the owner/author(s) .  \nACM ISBN 979-8-4007-2213-4/2026/11  \n[https://doi.org/10.1145/3767308.3836196](https://doi.org/10.1145/3767308.3836196)  \nCCS Concepts  \n• Computing methodologies → Robotic planning; Vision for robotics; • Computer systems organization → Robotics.  \nKeywords  \nAppliance manipulation planning, household robotics, multimodal foundation models  \nACM Reference Format:  \nYuxing Long, Lei Kang, Ziyan Yu, Yuzheng Gao, Bin Cheng, Jiyao Zhang, Xiaoqi Li, Haolin Yang, Dongjiang Li, Hui Shen, and Hao Dong. 2026. Scaling Manual-Grounded Appliance Manipulation with Data Synthesis and Unified Planning. In Proceedings of the 34th ACM International Conference on Multimedia (MM’26), November 10–14, 2026, Rio de Janeiro, Brazil. ACM, New York, NY, USA, 10 pages. [https://doi.org/10.1145/3767308.3836196](https://doi.org/10.1145/3767308.3836196)  \n1 Introduction  \nEnabling robots to operate household appliances is a key prerequisite for practical home autonomy. Appliance manipulation is substantially harder than generic pick-and-place: appliances vary widely in interface layout, operation logic, and state transitions, and successful execution requires long-horizon, state-dependent procedures that must remain robust to disturbances.  \nHumans address this by consulting manuals—an approach that naturally decomposes the task into three sub-c","cbCaipuFBfOGBOeB","https://ap.wps.com/l/cbCaipuFBfOGBOeB","pdf",4313262,10,"English","# Introduction\n## Problem: manual-grounded appliance manipulation\n## Approach: MAGE and Hierarchical Appliance Graph\n## Contributions: UseAppliance dataset and AppliancePlan model","[{\"question\":\"Why do existing large models struggle with appliance manipulation planning?\",\"answer\":\"They require state-dependent, long-horizon procedures that remain robust under disturbances, but there is no sufficiently diverse, task-oriented dataset tailored to manual-grounded manipulation.\"},{\"question\":\"What is MAGE and how does it generate training data?\",\"answer\":\"MAGE is a scalable data synthesis pipeline that uses a Hierarchical Appliance Graph to represent manuals and generates grounding, long-horizon planning, and closed-loop recovery data via graph traversal, disturbance injection, and human verification.\"},{\"question\":\"What are UseAppliance and AppliancePlan in this work?\",\"answer\":\"UseAppliance is a large-scale dataset built from appliance manuals across 22 categories, containing part annotations, manipulation tasks, and closed-loop steps. AppliancePlan is an end-to-end model for manual-grounded appliance manipulation planning trained on UseAppliance.\"}]","Scaling Manual-Grounded Appliance Manipulation with Data Synthesis and Unified Planning | PDF",1789227392]