2018-10-23-kahn18a.md

title	abstract	keywords	layout	series	id	month	tex_title	firstpage	lastpage	page	order	cycles	bibtex_author	author	date	address	publisher	container-title	volume	genre	issued	pdf	extras
Composable Action-Conditioned Predictors: Flexible Off-Policy Learning for Robot Navigation	A general-purpose intelligent robot must be able to learn autonomously and be able to accomplish multiple tasks in order to be deployed in the real world. However, standard reinforcement learning approaches learn separate task-specific policies and assume the reward function for each task is known a priori. We propose a framework that learns event cues from off-policy data, and can flexibly combine these event cues at test time to accomplish different tasks. These event cue labels are not assumed to be known a priori, but are instead labeled using learned models, such as computer vision detectors, and then “backed up” in time using an action-conditioned predictive model. We show that a simulated robotic car and a real-world RC car can gather data and train fully autonomously without any human-provided labels beyond those needed to train the detectors, and then at test-time be able to accomplish a variety of different tasks. Videos of the experiments and code can be found at github.com/gkahn13/CAPs	multi-task learning, rewards, prediction, robot navigation	inproceedings	Proceedings of Machine Learning Research	kahn18a	0	Composable Action-Conditioned Predictors: Flexible Off-Policy Learning for Robot Navigation	806	816	806-816	806	false	Kahn, Gregory and Villaflor, Adam and Abbeel, Pieter and Levine, Sergey	given family Gregory Kahn given family Adam Villaflor given family Pieter Abbeel given family Sergey Levine	2018-10-23		PMLR	Proceedings of The 2nd Conference on Robot Learning	87	inproceedings	date-parts 2018 10 23	http://proceedings.mlr.press/v87/kahn18a/kahn18a.pdf	label link Source code https://github.com/gkahn13/CAPs