[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-85378-en":3,"doc-seo-85378-105":29,"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":20,"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":13,"seo_description":14,"update_tm":27,"read_time":28},85378,7971461741311,"Ophelia","https://ap-avatar.wpscdn.com/avatar/74000253aff267980c6?x-image-process=image/resize,m_fixed,w_180,h_180&k=1779345379180704826",8,"Research & Report","A Minimalist Retargeting-Guided Reinforcement Learning Recipe for Dexterous Manipulation","A minimalist retargeting-guided reinforcement learning pipeline, REGRIND, learns dexterous manipulation policies from a single human demonstration for contact-rich tool use. The method retargets human hand-object motion to a robot reference that preserves spatial hand-object and contact relationships, then trains a residual RL policy in simulation to track object-centric keypoints along the reference. After zero-shot transfer, policies perform fluid, human-like actions on two multi-fingered robot hands across scissors and screwdriver tasks, with hardware experiments analyzing sim-to-real transfer factors and offering practical guidance.","arXiv :2607 . 11874v1 [ cs .RO] 13 Jul 2026  \nA Minimalist Retargeting-Guided Reinforcement Learning Recipe for Dexterous Manipulation  \nYunhai Feng 1 Natalie Leung 1 Jiaxuan Wang 1 Lujie Yang2 Haozhi Qi2 Preston Culbertson 1  \n1 Cornell University 2Amazon FAR  \n(a)  \n (b)  \n (c) (d)  \nFigure 1: Dexterous Tool Use Tasks. We present a minimalist retargeting-guided RL pipeline for learning dexterous manipulation from a single human demonstration. We evaluate our method on four contact-rich task-hand settings across two robot hands: (a) LEAP-Scissors, (b) WUJI-Screwdriver,(c) LEAP-Screwdriver, and (d) WUJI-Scissors.  \nAbstract: Recent work in humanoid whole-body control has found success with  \na simple recipe: retarget human motion to robot kinematic references, then train policies via reinforcement learning (RL) to track them. But how does this recipe transfer to dexterous manipulation? The answer is not obvious, as manipulation involves complex, contact-rich dynamics and requires delicate regulation of contact modes and forces. We present REGRIND, a minimalist retargeting-guided RL pipeline that learns dexterous manipulation policies from a single human demonstration. REGRIND retargets human hand-object motion to a robot reference that preserves hand-object spatial and contact relationships, trains a residual RL policy in simulation to track object-centric keypoints along that reference, and transfers the resulting policy zero-shot to hardware with careful system identification. The resulting policies produce fluid, human-like behavior on two different multi-fingered hands across contact-rich tool-use tasks, including operating a pair of scissors and turning a screwdriver. Through systematic hardware experiments, we identify and analyze the key factors that govern sim-to-real transfer in dexterous manipulation, offering practical guidance for retargeting-based learning in contact-rich settings. Videos and code are available at [https://yunhaifeng.com/REGRIND](https://yunhaifeng.com/REGRIND).  \nKeywords: Dexterous Manipulation, Motion Retargeting, Sim-to-Real  \n1 Introduction  \nHumanoid robots and anthropomorphic hands share the human body’s kinematic structure, opening the possibility of learning skills directly from the vast amount of human motion data. In humanoid whole-body control, this has led to a simple and increasingly effective recipe: retarget a human motion reference to the robot, train a policy in simulation to track the retargeted reference, and deploy the policy on hardware [1, 2] . Dexterous manipulation, however, has largely followed a different path. Current manipulation skills are often learned from teleoperated demonstrations collected directly on real robots [3, 4, 5], with limited use of either human motion data or massive simulation.  \nWhile recent work has explored using humanoid-style retargeting pipelines for dexterous hands, results have largely been confined to simulation or open-loop replays of simulation motions [6, 7] . One hypothesis for this limited real-world success is that simple kinematic retargeting, while closely matching the human hand pose, ignores the hand-object interaction and produces physically implausible trajectories that are a poor reference for downstream RL. This raises a natural question: does preserving interactions during retargeting improve RL for contact-rich dexterous manipulation?  \nIn this work, we answer this question by presenting REGRIND: REtargeting-Guided ReINforcement learning for Dexterous manipulation, a minimalist pipeline for learning contact-rich dexterous manipulation skills from a single human demonstration. We first retarget human hand-object motion to an interaction-preserving robot reference that maintains the spatial relationships between the hand and the object. To mitigate the scarcity of multi-fingered dexterous manipulation data, we dynamically augment the retargeted reference with perturbed initial configurations to provide broader coverage of the ","cbCaijS2I2NIl721","https://ap.wps.com/l/cbCaijS2I2NIl721","pdf",13425173,1,19,"English","en",105,"# Abstract\n# Introduction\n# Related Work","[{\"question\":\"What is REGRIND and what problem does it address?\",\"answer\":\"REGRIND is a minimalist retargeting-guided RL pipeline for learning contact-rich dexterous manipulation from a single human demonstration. It targets the challenge of transferring retargeting-based recipes to manipulation where hand-object interactions and contact dynamics are critical.\"},{\"question\":\"How does REGRIND retarget demonstrations for dexterous tool-use tasks?\",\"answer\":\"REGRIND retargets human hand-object motion to a robot reference that preserves spatial hand-object relationships and contact structure. It also uses dynamic augmentation of the retargeted reference via perturbed initial configurations to cover more of the task space.\"},{\"question\":\"How is the learned policy transferred from simulation to real hardware?\",\"answer\":\"The residual RL policy is trained in simulation to track object-centric keypoints along the interaction-preserving reference. Transfer is performed zero-shot to hardware with careful system identification, and hardware experiments analyze the key sim-to-real factors.\"}]",1784202994,48,{"code":4,"msg":30,"data":31},"ok",{"site_id":24,"language":23,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":27},"a-minimalist-retargeting-guided-reinforcement-learning-recipe-for-dexterous-manipulation","",{"@graph":35,"@context":85},[36,53,68],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/a-minimalist-retargeting-guided-reinforcement-learning-recipe-for-dexterous-manipulation/85378/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-17","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What is REGRIND and what problem does it address?","Question",{"text":75,"@type":76},"REGRIND is a minimalist retargeting-guided RL pipeline for learning contact-rich dexterous manipulation from a single human demonstration. It targets the challenge of transferring retargeting-based recipes to manipulation where hand-object interactions and contact dynamics are critical.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does REGRIND retarget demonstrations for dexterous tool-use tasks?",{"text":80,"@type":76},"REGRIND retargets human hand-object motion to a robot reference that preserves spatial hand-object relationships and contact structure. It also uses dynamic augmentation of the retargeted reference via perturbed initial configurations to cover more of the task space.",{"name":82,"@type":73,"acceptedAnswer":83},"How is the learned policy transferred from simulation to real hardware?",{"text":84,"@type":76},"The residual RL policy is trained in simulation to track object-centric keypoints along the interaction-preserving reference. Transfer is performed zero-shot to hardware with careful system identification, and hardware experiments analyze the key sim-to-real factors.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"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":45,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":45,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":45,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":45,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":45,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":45,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":45,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":21,"doc_module":4,"doc_module_name":45,"category_name":136,"show_sort_weight":106,"slug":137},"General","general"]