[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-82582-en":3,"doc-seo-82582-105":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},82582,34359740700684,"Finn","https://ap-avatar.wpscdn.com/avatar/1f400023980c374ae676?_k=1777273430885731487",8,"Research & Report","FAR Failure-Aware Retry for Test-Time Recovery and Continual Policy Improvement","Robot manipulation policies inevitably fail when deployed in real environments, and naive retries tend to repeat the same mistake while offline demonstrations provide little coverage of recovery states. Failure-Aware Retry (FAR) enables test-time learning from failures to adapt robot behavior and complete tasks autonomously. FAR uses Failure-Contrastive Preference Adaptation to steer away from failure-inducing actions and lightweight action perturbations to encourage local exploration, then incorporates successful recovery trajectories for continual policy improvement. Experiments show substantial robustness and success-rate gains in simulation and real-world manipulation, improving data efficiency under budgeted continual learning.","arXiv :2607 .0 1 1 1 1v 1 [ cs .RO] 1 Jul 2026  \nFAR: Failure-Aware Retry for Test-Time Recovery and Continual Policy Improvement  \nHaoran Hao Shahram Najam Syed Jeffrey Ichnowski Jeff Schneider  \nCarnegie Mellon University  \nAbstract: Robot policies inevitably encounter failures when deployed in real environments. Naive retries often repeat the same mistakes, while many existing recovery methods rely on human intervention. In this paper, we propose FailureAware Retry (FAR), a framework that enables robots to learn from previous failuresat test time, adapt their behavior accordingly, and eventually complete the task autonomously. FAR combines Failure-Contrastive Preference Adaptation, which constructs preference learning data from failures to steer the policy away from previously unsuccessful behaviors, with lightweight action perturbations during retries to encourage local exploration. We further incorporate successful recovery trajectories into a training loop for continual policy improvement. Experiments in both simulation and real-world manipulation tasks show that FAR substantially improves success rates and robustness, with average gains of 17.6% over the standard diffusion policy in simulation and 11.7% in the real world. In addition, FAR significantly improves data efficiency under both reset and timestep budgets during continual policy improvement by exploiting informative failure cases.  \nKeywords: Robot Manipulation, Failure Recovery, Test-time Adaptation  \n1 Introduction  \nRobot policies trained on offline expert demonstrations have achieved strong performance on a wide range of manipulation tasks [1, 2, 3, 4] . However, when deployed in real environments, these policies inevitably encounter failures, such as missing the target during grasping or dropping an object during execution. Recovering from such failures is challenging because these states are rarely covered by offline demonstrations and are often out-of-distribution for the pretrained policy. As a result, simply re-executing the same policy often repeats the same mistake.  \nExisting methods often rely on human intervention to correct robot behavior and collect additional data for future training [5, 6, 7], including DAgger and related human-in-the-loop methods [8, 9, 10, 11] . While effective, these approaches require substantial human effort. This raises the question of whether a robot can learn from failures, adapt its behavior at test time, and recover autonomously.  \nIn this work, we propose a framework for learning from failure and adapting robot behavior at test time to improve retry-based recovery [12, 13] . Effective retries face two key challenges: the policy tends to repeat the same mistake after failure, and its action distribution comes from offline demonstrations, limiting exploration in out-of-distribution recovery states. Our approach addresses these challenges with Failure-Contrastive Preference Adaptation (FCPA), which constructs preference learning signals from failure experience to steer the policy away from previously unsuccessful behaviors, and action perturbations during retries to encourage local exploration. Together, they improve retry success and generate diverse recovery trajectories for continual policy improvement.  \nThese recovery trajectories are valuable because they reveal the limitations and failure boundaries of the current policy, providing supervision that is absent from offline training. We therefore use them to further improve the policy. Starting from a policy pretrained on offline demonstrations, the robot interacts with the environment and makes multiple attempts on each task using FAR. The resulting  \nFigure 1: Overall Framework of FAR. After a failure, FAR identifies failure-inducing actions using value estimation, then updates the policy with both failure examples and alternative positive examples. The collected trajectories are added to the replay buffer for continual policy improvement.  \ntrajectories are inco","cbCaivAojFySssO7","https://ap.wps.com/l/cbCaivAojFySssO7","pdf",1258853,1,22,"English","en",105,"# Introduction\n# Related Work","[{\"question\":\"What problem does FAR address during test-time robot recovery?\",\"answer\":\"FAR targets two failure issues: repeated mistakes after a failure and limited exploration because action choices are constrained by policies learned from offline demonstrations.\"},{\"question\":\"How does Failure-Contrastive Preference Adaptation help the robot after failures?\",\"answer\":\"FCPA builds preference signals from failure experience to steer the policy away from previously unsuccessful behaviors, using both failure cases and alternative positive examples.\"},{\"question\":\"How does FAR improve continual policy performance beyond single retries?\",\"answer\":\"FAR adds collected recovery trajectories into a replay buffer and uses them for online fine-tuning, enabling continual policy improvement and better data efficiency under reset and timestep budgets.\"}]",1784181642,55,{"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},"far-failure-aware-retry-for-test-time-recovery-and-continual-policy-improvement","",{"@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/far-failure-aware-retry-for-test-time-recovery-and-continual-policy-improvement/82582/",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 problem does FAR address during test-time robot recovery?","Question",{"text":75,"@type":76},"FAR targets two failure issues: repeated mistakes after a failure and limited exploration because action choices are constrained by policies learned from offline demonstrations.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does Failure-Contrastive Preference Adaptation help the robot after failures?",{"text":80,"@type":76},"FCPA builds preference signals from failure experience to steer the policy away from previously unsuccessful behaviors, using both failure cases and alternative positive examples.",{"name":82,"@type":73,"acceptedAnswer":83},"How does FAR improve continual policy performance beyond single retries?",{"text":84,"@type":76},"FAR adds collected recovery trajectories into a replay buffer and uses them for online fine-tuning, enabling continual policy improvement and better data efficiency under reset and timestep budgets.","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":136,"doc_module":4,"doc_module_name":45,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]