{"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/using-simulation-and-domain-adaptation-to","title":"Using Simulation and Domain Adaptation to Improve Efficiency of Deep Robotic Grasping","arxiv_id":"1709.07857","date":"2017-09-22","proceeding":null,"authors":["Konstantinos Bousmalis","Alex Irpan","Paul Wohlhart","Yunfei Bai","Matthew Kelcey","Mrinal Kalakrishnan","Laura Downs","Julian Ibarz","Peter Pastor","Kurt Konolige","Sergey Levine","Vincent Vanhoucke"],"abstract":"Instrumenting and collecting annotated visual grasping datasets to train\nmodern machine learning algorithms can be extremely time-consuming and\nexpensive. An appealing alternative is to use off-the-shelf simulators to\nrender synthetic data for which ground-truth annotations are generated\nautomatically. Unfortunately, models trained purely on simulated data often\nfail to generalize to the real world. We study how randomized simulated\nenvironments and domain adaptation methods can be extended to train a grasping\nsystem to grasp novel objects from raw monocular RGB images. We extensively\nevaluate our approaches with a total of more than 25,000 physical test grasps,\nstudying a range of simulation conditions and domain adaptation methods,\nincluding a novel extension of pixel-level domain adaptation that we term the\nGraspGAN. We show that, by using synthetic data and domain adaptation, we are\nable to reduce the number of real-world samples needed to achieve a given level\nof performance by up to 50 times, using only randomly generated simulated\nobjects. We also show that by using only unlabeled real-world data and our\nGraspGAN methodology, we obtain real-world grasping performance without any\nreal-world labels that is similar to that achieved with 939,777 labeled\nreal-world samples.","url_abs":"http://arxiv.org/abs/1709.07857v2","url_pdf":"http://arxiv.org/pdf/1709.07857v2.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":"using-simulation-and-domain-adaptation-to","repo_url":"https://github.com/sabeaussan/ROS_Unity","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}}],"tasks":[{"task_slug":"domain-adaptation","task_name":"Domain Adaptation"},{"task_slug":"industrial-robots","task_name":"Industrial Robots"},{"task_slug":"robotic-grasping","task_name":"Robotic Grasping"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1709.07857","atlas_url":"https://app.syntology.ai/?focus=1709.07857","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}