Carlos Florensa,
David Held,
Markus Wulfmeier,
Michael Zhang,
Pieter Abbeel
;
Proceedings of the 1st Annual Conference on Robot Learning, PMLR 78:482-495, 2017.
Abstract
Many relevant tasks require an agent to reach a certain state, or to manipulate objects into a desired configuration. For example, we might want a robot to align and assemble a gear onto an axle or insert and turn a key in a lock. These goal-oriented tasks present a considerable challenge for reinforcement learning, since their natural reward function is sparse and prohibitive amounts of exploration are required to reach the goal and receive some learning signal. Past approaches tackle these problems by exploiting expert demonstrations or by manually designing a task-specific reward shaping function to guide the learning agent. Instead, we propose a method to learn these tasks without requiring any prior knowledge other than obtaining a single state in which the task is achieved. The robot is trained in “reverse", gradually learning to reach the goal from a set of starting positions increasingly far from the goal. Our method automatically generates a curriculum of starting positions that adapts to the agent’s performance, leading to efficient training on goal-oriented tasks. We demonstrate our approach on difficult simulated navigation and fine-grained manipulation problems, not solvable by state-of-the-art reinforcement learning methods.
@InProceedings{pmlr-v78-florensa17a,
title = {Reverse Curriculum Generation for Reinforcement Learning},
author = {Carlos Florensa and David Held and Markus Wulfmeier and Michael Zhang and Pieter Abbeel},
booktitle = {Proceedings of the 1st Annual Conference on Robot Learning},
pages = {482--495},
year = {2017},
editor = {Sergey Levine and Vincent Vanhoucke and Ken Goldberg},
volume = {78},
series = {Proceedings of Machine Learning Research},
address = {},
month = {13--15 Nov},
publisher = {PMLR},
pdf = {http://proceedings.mlr.press/v78/florensa17a/florensa17a.pdf},
url = {http://proceedings.mlr.press/v78/florensa17a.html},
abstract = {Many relevant tasks require an agent to reach a certain state, or to manipulate objects into a desired configuration. For example, we might want a robot to align and assemble a gear onto an axle or insert and turn a key in a lock. These goal-oriented tasks present a considerable challenge for reinforcement learning, since their natural reward function is sparse and prohibitive amounts of exploration are required to reach the goal and receive some learning signal. Past approaches tackle these problems by exploiting expert demonstrations or by manually designing a task-specific reward shaping function to guide the learning agent. Instead, we propose a method to learn these tasks without requiring any prior knowledge other than obtaining a single state in which the task is achieved. The robot is trained in “reverse", gradually learning to reach the goal from a set of starting positions increasingly far from the goal. Our method automatically generates a curriculum of starting positions that adapts to the agent’s performance, leading to efficient training on goal-oriented tasks. We demonstrate our approach on difficult simulated navigation and fine-grained manipulation problems, not solvable by state-of-the-art reinforcement learning methods.}
}
%0 Conference Paper
%T Reverse Curriculum Generation for Reinforcement Learning
%A Carlos Florensa
%A David Held
%A Markus Wulfmeier
%A Michael Zhang
%A Pieter Abbeel
%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-florensa17a
%I PMLR
%J Proceedings of Machine Learning Research
%P 482--495
%U http://proceedings.mlr.press
%V 78
%W PMLR
%X Many relevant tasks require an agent to reach a certain state, or to manipulate objects into a desired configuration. For example, we might want a robot to align and assemble a gear onto an axle or insert and turn a key in a lock. These goal-oriented tasks present a considerable challenge for reinforcement learning, since their natural reward function is sparse and prohibitive amounts of exploration are required to reach the goal and receive some learning signal. Past approaches tackle these problems by exploiting expert demonstrations or by manually designing a task-specific reward shaping function to guide the learning agent. Instead, we propose a method to learn these tasks without requiring any prior knowledge other than obtaining a single state in which the task is achieved. The robot is trained in “reverse", gradually learning to reach the goal from a set of starting positions increasingly far from the goal. Our method automatically generates a curriculum of starting positions that adapts to the agent’s performance, leading to efficient training on goal-oriented tasks. We demonstrate our approach on difficult simulated navigation and fine-grained manipulation problems, not solvable by state-of-the-art reinforcement learning methods.
Florensa, C., Held, D., Wulfmeier, M., Zhang, M. & Abbeel, P.. (2017). Reverse Curriculum Generation for Reinforcement Learning. Proceedings of the 1st Annual Conference on Robot Learning, in PMLR 78:482-495
This site last compiled Fri, 22 Dec 2017 01:53:11 +0000