Show simple item record

dc.contributor.authorArulkumaran, Kai
dc.contributor.authorAshley, Dylan R.
dc.contributor.authorSchmidhuber, Juergen
dc.contributor.authorSrivastava, Rupesh K.
dc.date.accessioned2022-05-16T11:58:52Z
dc.date.available2022-05-16T11:58:52Z
dc.date.issued2022-02-24
dc.identifier.urihttp://hdl.handle.net/10754/677960
dc.description.abstractUpside down reinforcement learning (UDRL) flips the conventional use of the return in the objective function in RL upside down, by taking returns as input and predicting actions. UDRL is based purely on supervised learning, and bypasses some prominent issues in RL: bootstrapping, off-policy corrections, and discount factors. While previous work with UDRL demonstrated it in a traditional online RL setting, here we show that this single algorithm can also work in the imitation learning and offline RL settings, be extended to the goal-conditioned RL setting, and even the meta-RL setting. With a general agent architecture, a single UDRL agent can learn across all paradigms.
dc.description.sponsorshipThis work was partially supported by the European Research Council (ERC, Advanced Grant Number 742870).
dc.publisherarXiv
dc.relation.urlhttps://arxiv.org/pdf/2202.11960.pdf
dc.rightsArchived with thanks to arXiv
dc.titleAll You Need Is Supervised Learning: From Imitation Learning to Meta-RL With Upside Down RL
dc.typePreprint
dc.contributor.departmentComputer Science Program
dc.contributor.departmentComputer, Electrical and Mathematical Science and Engineering (CEMSE) Division
dc.eprint.versionPre-print
dc.contributor.institutionThe Swiss AI Lab IDSIA, Lugano, Switzerland
dc.contributor.institutionUniversità della Svizzera Italiana (USI), Lugano, Switzerland
dc.contributor.institutionScuola Universitaria Professionale della Svizzera Italiana (SUPSI), Lugano, Switzerland
dc.contributor.institutionNNAISENSE, Lugano, Switzerland
dc.identifier.arxivid2202.11960
kaust.personSchmidhuber, Juergen
refterms.dateFOA2022-05-16T11:59:53Z


Files in this item

Thumbnail
Name:
2202.11960.pdf
Size:
419.7Kb
Format:
PDF
Description:
Preprint

This item appears in the following Collection(s)

Show simple item record