Show simple item record

dc.contributor.author
Ackermann, Johannes
dc.contributor.author
Richter, Oliver
dc.contributor.author
Wattenhofer, Roger
dc.date.accessioned
2021-03-25T13:55:47Z
dc.date.available
2021-01-11T08:55:52Z
dc.date.available
2021-03-25T13:55:47Z
dc.date.issued
2020-12
dc.identifier.uri
http://hdl.handle.net/20.500.11850/461024
dc.description.abstract
Meta-learning, transfer learning and multi-task learning have recently laid a path towards more generally applicable reinforcement learning agents that are not limited to a single task. However, most existing approaches implicitly assume a uniform similarity between tasks. We argue that this assumption is limiting in settings where the relationship between tasks is unknown a-priori. In this work, we propose a general approach to automatically cluster together similar tasks during training. Our method, inspired by the expectation maximization algorithm, succeeds at finding clusters of related tasks and uses these to improve sample complexity. In the expectation step, we evaluate the performance of a set of policies on all tasks and assign each task to the best performing policy. In the maximization step, each policy trains by sampling tasks from its assigned set. This method is intuitive, simple to implement and orthogonal to other multi-task learning algorithms. We show the generality of our approach by evaluating on simple discrete and continuous control tasks, as well as complex bipedal walker tasks and Atari games. Results show improvements in sample complexity as well as a more general applicability when compared to other approaches.
en_US
dc.language.iso
en
en_US
dc.publisher
Deep RL Workshop
en_US
dc.title
Unsupervised Task Clustering for Multi-Task Reinforcement Learning
en_US
dc.type
Conference Paper
ethz.book.title
Deep Reinforcement Learning Workshop, NeurIPS 2020. Accepted Papers
en_US
ethz.size
16 p.
en_US
ethz.event
Deep Reinforcement Learning Workshop at the 34th Conference on Neural Information Processing Systems (NeurIPS 2020) (virtual)
en_US
ethz.event.location
Vancouver, Canada
en_US
ethz.event.date
December 11, 2020
en_US
ethz.notes
Due to the Coronavirus (COVID-19) the workshop was conducted virtually.
en_US
ethz.publication.place
s.l.
en_US
ethz.publication.status
published
en_US
ethz.leitzahl
ETH Zürich::00002 - ETH Zürich::00012 - Lehre und Forschung::00007 - Departemente::02140 - Dep. Inf.technologie und Elektrotechnik / Dep. of Inform.Technol. Electrical Eng.::02640 - Inst. f. Technische Informatik und Komm. / Computer Eng. and Networks Lab.::03604 - Wattenhofer, Roger / Wattenhofer, Roger
en_US
ethz.leitzahl.certified
ETH Zürich::00002 - ETH Zürich::00012 - Lehre und Forschung::00007 - Departemente::02140 - Dep. Inf.technologie und Elektrotechnik / Dep. of Inform.Technol. Electrical Eng.::02640 - Inst. f. Technische Informatik und Komm. / Computer Eng. and Networks Lab.::03604 - Wattenhofer, Roger / Wattenhofer, Roger
en_US
ethz.identifier.url
https://sites.google.com/view/deep-rl-workshop-neurips2020/home
ethz.date.deposited
2021-01-11T08:55:59Z
ethz.source
FORM
ethz.eth
yes
en_US
ethz.availability
Metadata only
en_US
ethz.rosetta.installDate
2021-03-25T13:55:57Z
ethz.rosetta.lastUpdated
2021-03-25T13:55:57Z
ethz.rosetta.versionExported
true
ethz.COinS
ctx_ver=Z39.88-2004&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.atitle=Unsupervised%20Task%20Clustering%20for%20Multi-Task%20Reinforcement%20Learning&rft.date=2020-12&rft.au=Ackermann,%20Johannes&Richter,%20Oliver&Wattenhofer,%20Roger&rft.genre=proceeding&rft.btitle=Deep%20Reinforcement%20Learning%20Workshop,%20NeurIPS%202020.%20Accepted%20Papers
 Search print copy at ETH Library

Files in this item

FilesSizeFormatOpen in viewer

There are no files associated with this item.

Publication type

Show simple item record