flâneur

[2310.17555] Interactive Robot Learning from Verbal Correction

ar5iv.labs.arxiv.org · 9,521 words · saved by 1 readers

The ability to learn and refine behavior after deployment has become ever more important for robots as we design them to operate in unstructured environments like households. In this work, we design a new learning syst…

Interactive Robot Learning from Verbal Correction Huihan Liu 1,2 , Alice Chen 1 , Yuke Zhu 2 , Adith Swaminathan 1 , Andrey Kolobov 1 , Ching-An Cheng 1 *This work was partially done during Huihan Liu’s internship at Microsoft Research. Correspondence: huihanl@utexas.edu 1 Microsoft Research, Redmond, WA 98052, USA {adswamin,akolobov,chinganc}@microsoft.com 2 The University of Texas at Austin, Austin, TX 78712, USA {huihanl,yukez}@utexas.edu Abstract The ability to learn and refine behavior after deployment has become ever more important for robots as we design them to operate in unstructured

related reading