One-Shot Visual Imitation Learning via Meta-Learning

Chelsea Finn, Tianhe Yu, Tianhao Zhang, Pieter Abbeel, Sergey Levine
Proceedings of the 1st Annual Conference on Robot Learning, PMLR 78:357-368, 2017.

Abstract

In order for a robot to be a generalist that can perform a wide range of jobs, it must be able to acquire a wide variety of skills quickly and efficiently in complex unstructured environments. High-capacity models such as deep neural networks can enable a robot to represent complex skills, but learning each skill from scratch then becomes infeasible. In this work, we present a meta-imitation learning method that enables a robot to learn how to learn more efficiently, allowing it to acquire new skills from just a single demonstration. Unlike prior methods for one-shot imitation, our method can scale to raw pixel inputs and requires data from significantly fewer prior tasks for effective learning of new skills. Our experiments on both simulated and real robot platforms demonstrate the ability to learn new tasks, end-to-end, from a single visual demonstration.

Cite this Paper

BibTeX
@InProceedings{pmlr-v78-finn17a, title = {One-Shot Visual Imitation Learning via Meta-Learning}, author = {Finn, Chelsea and Yu, Tianhe and Zhang, Tianhao and Abbeel, Pieter and Levine, Sergey}, booktitle = {Proceedings of the 1st Annual Conference on Robot Learning}, pages = {357--368}, year = {2017}, editor = {Levine, Sergey and Vanhoucke, Vincent and Goldberg, Ken}, volume = {78}, series = {Proceedings of Machine Learning Research}, month = {13--15 Nov}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v78/finn17a/finn17a.pdf}, url = {https://proceedings.mlr.press/v78/finn17a.html}, abstract = {In order for a robot to be a generalist that can perform a wide range of jobs, it must be able to acquire a wide variety of skills quickly and efficiently in complex unstructured environments. High-capacity models such as deep neural networks can enable a robot to represent complex skills, but learning each skill from scratch then becomes infeasible. In this work, we present a meta-imitation learning method that enables a robot to learn how to learn more efficiently, allowing it to acquire new skills from just a single demonstration. Unlike prior methods for one-shot imitation, our method can scale to raw pixel inputs and requires data from significantly fewer prior tasks for effective learning of new skills. Our experiments on both simulated and real robot platforms demonstrate the ability to learn new tasks, end-to-end, from a single visual demonstration.} }
Endnote
%0 Conference Paper %T One-Shot Visual Imitation Learning via Meta-Learning %A Chelsea Finn %A Tianhe Yu %A Tianhao Zhang %A Pieter Abbeel %A Sergey Levine %B Proceedings of the 1st Annual Conference on Robot Learning %C Proceedings of Machine Learning Research %D 2017 %E Sergey Levine %E Vincent Vanhoucke %E Ken Goldberg %F pmlr-v78-finn17a %I PMLR %P 357--368 %U https://proceedings.mlr.press/v78/finn17a.html %V 78 %X In order for a robot to be a generalist that can perform a wide range of jobs, it must be able to acquire a wide variety of skills quickly and efficiently in complex unstructured environments. High-capacity models such as deep neural networks can enable a robot to represent complex skills, but learning each skill from scratch then becomes infeasible. In this work, we present a meta-imitation learning method that enables a robot to learn how to learn more efficiently, allowing it to acquire new skills from just a single demonstration. Unlike prior methods for one-shot imitation, our method can scale to raw pixel inputs and requires data from significantly fewer prior tasks for effective learning of new skills. Our experiments on both simulated and real robot platforms demonstrate the ability to learn new tasks, end-to-end, from a single visual demonstration.
APA
Finn, C., Yu, T., Zhang, T., Abbeel, P. & Levine, S.. (2017). One-Shot Visual Imitation Learning via Meta-Learning. Proceedings of the 1st Annual Conference on Robot Learning, in Proceedings of Machine Learning Research 78:357-368 Available from https://proceedings.mlr.press/v78/finn17a.html.

Related Material