Timezone: »

 
Spotlight
Translating Robot Skills: Learning Unsupervised Skill Correspondences Across Robots
Tanmay Shankar · Yixin Lin · Aravind Rajeswaran · Vikash Kumar · Stuart Anderson · Jean Oh

Wed Jul 20 03:00 PM -- 03:05 PM (PDT) @ Hall G
in Reinforcement Learning »

In this paper, we explore how we can endow robots with the ability to learn correspondences between their own skills, and those of morphologically different robots in different domains, in an entirely unsupervised manner. We make the insight that different morphological robots use similar task strategies to solve similar tasks. Based on this insight, we frame learning skill correspondences as a problem of matching distributions of sequences of skills across robots. We then present an unsupervised objective that encourages a learnt skill translation model to match these distributions across domains, inspired by recent advances in unsupervised machine translation. Our approach is able to learn semantically meaningful correspondences between skills across multiple robot-robot and human-robot domain pairs despite being completely unsupervised. Further, the learnt correspondences enable the transfer of task strategies across robots and domains. We present dynamic visualizations of our results at https://sites.google.com/view/translatingrobotskills/home.

Keywords: APP: Robotics · RL: Deep RL · MISC: Representation Learning · MISC: Unsupervised and Semi-supervised Learning · MISC: Transfer, Multitask and Meta-learning

Author Information

Tanmay Shankar (Carnegie Mellon University)
Yixin Lin (Meta AI)
Aravind Rajeswaran (Meta AI (FAIR))
Vikash Kumar (Univ. Of Washington)
Stuart Anderson (Facebook AI Research)
Jean Oh (Carnegie Mellon University)

Related Events (a corresponding poster, oral, or spotlight)

More from the Same Authors