[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-83835-en":3,"doc-seo-83835-105":30,"detail-sidebar-cat-0-en-105":92},{"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":20,"is_deleted":4,"is_public":21,"is_downloadable":21,"audit_status":21,"page_count":22,"language":23,"language_code":24,"site_id":25,"html_lang":24,"table_of_contents":26,"faqs":27,"seo_title":13,"seo_description":14,"update_tm":28,"read_time":29},83835,2336464648322,"Aria","https://ap-avatar.wpscdn.com/avatar/2200025388227c56fec?_k=1778556882303663488",8,"Research & Report","HUGS Guiding Unified Dexterous Grasp Synthesis Across Modes and Scales via Learned Human Priors","Dexterous grasping across object scales requires multiple contact modes, from two-finger pinches to bimanual grasps, yet conventional synthesis methods rely on manually chosen expected contacts and heuristic wrist initializations that cannot balance success rate and diversity. HUGS introduces a Human-prior-guided framework that learns an object-conditioned human prior from a compact dataset and guides force-closure-aware optimization. The learned prior adaptively proposes contact modes and wrist initializations, improving coverage–success trade-offs and enabling scalable grasp generation.","arXiv :2607 .04554v 1 [ cs .RO] 6 Jul 2026  \nHUGS: Guiding Unified Dexterous Grasp Synthesis Across Modes and Scales via Learned Human Priors  \nMingrui Yu* Yongpeng Jiang* Yongyi Jia Kangchen Lv Xiangjie Yan Li Huang Yi Ren Xiang Li†  \nTsinghua University  \n*Equal contribution. †Corresponding author.  \n[https://hugs-dex.github.io/](https://hugs-dex.github.io/)  \nFigure 1: Human priors guide scalable dexterous grasp synthesis across modes and scales. HUGS learns an object-conditioned human prior from a compact self-collected dataset to predict preferred contact modes and wrist initializations. Guided by this prior, force-closure-aware optimization synthesizes diverse and stable grasps ranging from two-finger pinches to bimanual grasps.  \nAbstract: Dexterous grasping across diverse object scales requires contact modes ranging from two-finger pinches to bimanual grasps. Existing dexterous grasp synthesis methods reduce the high-dimensional optimization space with manually designed expected contacts and initialization heuristics, which struggle to balance synthesis success rate and diversity. We present HUGS, a Human-prior-guided framework for Unified dexterous Grasp Synthesis across modes and scales. Instead of directly retargeting human demonstrations, HUGS learns an object-conditioned human prior that captures human grasp preferences and guides downstream forceclosure-aware optimization. The prior is trained on a compact self-collected human grasp dataset with 1.8K grasps over 304 objects, providing broad coverage of object scales and contact modes. During synthesis, HUGS adaptively proposes contact modes and wrist initializations, substantially improving the balance between contact-mode coverage and synthesis success rate over heuristic-based methods.  \nWith HUGS, we synthesize 3.2M robotic grasps over 157K scenes, spanning object half-diagonal lengths from 2 cm to 30 cm and modes from two-finger to bimanual  \ngrasps. Models trained on the synthesized dataset autonomously select appropriate contact modes in the real world, enabling grasping from screws to large boxes.  \nKeywords: Dexterous Grasping, Human Priors, Grasp Synthesis  \n1 Introduction  \nDexterous grasping is inherently multi-mode across object scales, from two-finger precision grasps to coordinated bimanual grasps, and the same object may admit different modes under different conditions. Training a generalizable grasp generation model for such diversity requires massive datasets, yet large-scale real-world collection is impractical, making scalable grasp synthesis essential.  \nDexterous grasp synthesis involves a vast search space due to the high degrees of freedom and multi-contact nature. Existing methods often constrain optimization with manually predefined contact modes, such as single-hand [1, 2] or dual-hand full-finger grasps [3, 4], and heuristic wrist pose initializations. However, these coarse rules fail to exploit object-specific information and struggle to balance synthesis efficiency and grasp diversity: loose heuristics lead to low-quality optimization and inefficient synthesis, while restrictive heuristics severely limit grasp diversity.  \nIndeed, anthropomorphic robotic hands are designed for human-like manipulation [5], making human grasps a natural source of knowledge for robots. Existing methods often retarget human grasps to robotic hands [6, 7, 8], converting each demonstration into a robot grasp for the same or highly similar object and thereby limiting scalable synthesis over large-scale, diverse object sets.  \nOur key insight is that we can use human demonstrations to learn generalized and reusable objectconditioned human priors, instead of directly retargeting each human grasp to a robotic hand in a one-to-one manner. By capturing human preferences for specific object geometry, the prior guides robotic grasp optimization with better global initializations and optimization targets for efficient convergence to higher-quality solutions. Such a ","cbCaidbTRyzLXOBK","https://ap.wps.com/l/cbCaidbTRyzLXOBK","pdf",12477960,5,1,30,"English","en",105,"# Introduction\n## Problem: multi-mode dexterous grasp synthesis\n## Limitations of heuristic and retargeting methods\n## Key insight: learned object-conditioned human priors\n## Proposed framework: HUGS and contributions","[{\"question\":\"What problem does HUGS address in dexterous grasp synthesis?\",\"answer\":\"HUGS targets the need to synthesize diverse dexterous grasps across object scales and contact modes while avoiding the trade-off failures of heuristic-based searches.\"},{\"question\":\"How does HUGS use human information without direct retargeting?\",\"answer\":\"HUGS learns an object-conditioned human prior that captures human grasp preferences and then guides force-closure-aware robotic optimization toward better contact-mode and wrist-initialization choices.\"},{\"question\":\"What dataset scale and outcomes are reported for HUGS?\",\"answer\":\"HUGS trains the prior on a compact self-collected set of 1.8K grasps over 304 objects and synthesizes 3.2M robotic grasps over 157K scenes, covering object scales from 2 cm to 30 cm and modes from two-finger to bimanual grasps.\"}]",1784190871,76,{"code":4,"msg":31,"data":32},"ok",{"site_id":25,"language":24,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":87,"head_meta":89,"extra_data":91,"updated_unix":28},"hugs-guiding-unified-dexterous-grasp-synthesis-across-modes-and-scales-via-learned-human-priors","",{"@graph":36,"@context":86},[37,54,69],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":21},"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/hugs-guiding-unified-dexterous-grasp-synthesis-across-modes-and-scales-via-learned-human-priors/83835/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":24,"description":14,"dateModified":62,"datePublished":63,"encodingFormat":61,"isAccessibleForFree":64,"interactionStatistic":65},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-07-27","2026-07-16",true,{"@type":66,"interactionType":67,"userInteractionCount":20},"InteractionCounter",{"@type":68},"ViewAction",{"@type":70,"mainEntity":71},"FAQPage",[72,78,82],{"name":73,"@type":74,"acceptedAnswer":75},"What problem does HUGS address in dexterous grasp synthesis?","Question",{"text":76,"@type":77},"HUGS targets the need to synthesize diverse dexterous grasps across object scales and contact modes while avoiding the trade-off failures of heuristic-based searches.","Answer",{"name":79,"@type":74,"acceptedAnswer":80},"How does HUGS use human information without direct retargeting?",{"text":81,"@type":77},"HUGS learns an object-conditioned human prior that captures human grasp preferences and then guides force-closure-aware robotic optimization toward better contact-mode and wrist-initialization choices.",{"name":83,"@type":74,"acceptedAnswer":84},"What dataset scale and outcomes are reported for HUGS?",{"text":85,"@type":77},"HUGS trains the prior on a compact self-collected set of 1.8K grasps over 304 objects and synthesizes 3.2M robotic grasps over 157K scenes, covering object scales from 2 cm to 30 cm and modes from two-finger to bimanual grasps.","https://schema.org",{"og:url":52,"og:type":88,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":90,"canonical":52},"index,follow",{"doc_id":7,"site_id":25},{"code":4,"msg":5,"data":93},[94,98,102,106,110,115,120,122,127,130,134],{"id":21,"doc_module":4,"doc_module_name":46,"category_name":95,"show_sort_weight":96,"slug":97},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":99,"show_sort_weight":100,"slug":101},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":103,"show_sort_weight":104,"slug":105},"Exam",70,"exam",{"id":20,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},"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":22,"slug":121},"research-report",{"id":123,"doc_module":4,"doc_module_name":46,"category_name":124,"show_sort_weight":125,"slug":126},9,"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":46,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":46,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":46,"category_name":136,"show_sort_weight":20,"slug":137},19,"General","general"]