2018-10-23-cheng18a.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
Reinforcement Learning of Active Vision for Manipulating Objects under Occlusions	We consider artificial agents that learn to jointly control their gripper and camera in order to reinforcement learn manipulation policies in the presence of occlusions from distractor objects. Distractors often occlude the object of interest and cause it to disappear from the field of view. We propose hand/eye controllers that learn to move the camera to keep the object within the field of view and visible, in coordination to manipulating it to achieve the desired goal, e.g., pushing it to a target location. We incorporate structural biases of object-centric attention within our actor-critic architectures, which our experiments suggest to be a key for good performance. Our results further highlight the importance of curriculum with regards to environment difficulty. The resulting active vision / manipulation policies outperform static camera setups for a variety of cluttered environments.	Manipulation, Reinforcement learning, Control	inproceedings	Proceedings of Machine Learning Research	cheng18a	0	Reinforcement Learning of Active Vision for Manipulating Objects under Occlusions	422	431	422-431	422	false	Cheng, Ricson and Agarwal, Arpit and Fragkiadaki, Katerina	given family Ricson Cheng given family Arpit Agarwal given family Katerina Fragkiadaki	2018-10-23		PMLR	Proceedings of The 2nd Conference on Robot Learning	87	inproceedings	date-parts 2018 10 23	http://proceedings.mlr.press/v87/cheng18a/cheng18a.pdf	label link Source code https://github.com/ricsonc/ActiveVisionManipulation