{"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/adversarial-discriminative-sim-to-real","title":"Adversarial Discriminative Sim-to-real Transfer of Visuo-motor Policies","arxiv_id":"1709.05746","date":"2017-09-18","proceeding":null,"authors":["Fangyi Zhang","Jürgen Leitner","ZongYuan Ge","Michael Milford","Peter Corke"],"abstract":"Various approaches have been proposed to learn visuo-motor policies for\nreal-world robotic applications. One solution is first learning in simulation\nthen transferring to the real world. In the transfer, most existing approaches\nneed real-world images with labels. However, the labelling process is often\nexpensive or even impractical in many robotic applications. In this paper, we\npropose an adversarial discriminative sim-to-real transfer approach to reduce\nthe cost of labelling real data. The effectiveness of the approach is\ndemonstrated with modular networks in a table-top object reaching task where a\n7 DoF arm is controlled in velocity mode to reach a blue cuboid in clutter\nthrough visual observations. The adversarial transfer approach reduced the\nlabelled real data requirement by 50%. Policies can be transferred to real\nenvironments with only 93 labelled and 186 unlabelled real images. The\ntransferred visuo-motor policies are robust to novel (not seen in training)\nobjects in clutter and even a moving target, achieving a 97.8% success rate and\n1.8 cm control accuracy.","url_abs":"http://arxiv.org/abs/1709.05746v2","url_pdf":"http://arxiv.org/pdf/1709.05746v2.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":"adversarial-discriminative-sim-to-real","repo_url":"https://github.com/Fanleyrobot/ADT","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"torch","reach":null}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1709.05746","atlas_url":"https://app.syntology.ai/?focus=1709.05746","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}