{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/qt-opt-scalable-deep-reinforcement-learning","title":"QT-Opt: Scalable Deep Reinforcement Learning for Vision-Based Robotic Manipulation","arxiv_id":"1806.10293","date":"2018-06-27","proceeding":null,"authors":["Dmitry Kalashnikov","Alex Irpan","Peter Pastor","Julian Ibarz","Alexander Herzog","Eric Jang","Deirdre Quillen","Ethan Holly","Mrinal Kalakrishnan","Vincent Vanhoucke","Sergey Levine"],"abstract":"In this paper, we study the problem of learning vision-based dynamic\nmanipulation skills using a scalable reinforcement learning approach. We study\nthis problem in the context of grasping, a longstanding challenge in robotic\nmanipulation. In contrast to static learning behaviors that choose a grasp\npoint and then execute the desired grasp, our method enables closed-loop\nvision-based control, whereby the robot continuously updates its grasp strategy\nbased on the most recent observations to optimize long-horizon grasp success.\nTo that end, we introduce QT-Opt, a scalable self-supervised vision-based\nreinforcement learning framework that can leverage over 580k real-world grasp\nattempts to train a deep neural network Q-function with over 1.2M parameters to\nperform closed-loop, real-world grasping that generalizes to 96% grasp success\non unseen objects. Aside from attaining a very high success rate, our method\nexhibits behaviors that are quite distinct from more standard grasping systems:\nusing only RGB vision-based perception from an over-the-shoulder camera, our\nmethod automatically learns regrasping strategies, probes objects to find the\nmost effective grasps, learns to reposition objects and perform other\nnon-prehensile pre-grasp manipulations, and responds dynamically to\ndisturbances and perturbations.","url_abs":"http://arxiv.org/abs/1806.10293v3","url_pdf":"http://arxiv.org/pdf/1806.10293v3.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"qt-opt-scalable-deep-reinforcement-learning","repo_url":"https://github.com/hyecheol123/Summary_of_QT-Opt","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}}],"tasks":[{"task_slug":"deep-reinforcement-learning","task_name":"Deep Reinforcement Learning"},{"task_slug":"reinforcement-learning","task_name":"Reinforcement Learning"},{"task_slug":"reinforcement-learning-1","task_name":"Reinforcement Learning (RL)"},{"task_slug":"reinforcement-learning-2","task_name":"reinforcement-learning"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1806.10293","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}