Show simple item record

dc.contributor.authorJacob, D.
dc.contributor.authorPolani, D.
dc.contributor.authorNehaniv, C.L.
dc.date.accessioned2008-03-04T12:10:25Z
dc.date.available2008-03-04T12:10:25Z
dc.date.issued2005
dc.identifier.citationJacob , D , Polani , D & Nehaniv , C L 2005 , ' Inferring dependencies in Embodiment-based modular reinforcement learning ' , TAROS , vol. 2005 , pp. 103-110 .
dc.identifier.otherPURE: 94175
dc.identifier.otherPURE UUID: 3971214c-fd03-42bf-b595-26dbea14cbe5
dc.identifier.otherdspace: 2299/1728
dc.identifier.otherORCID: /0000-0002-3233-5847/work/86098077
dc.identifier.urihttp://hdl.handle.net/2299/1728
dc.description.abstractThe state-spaces needed to describe realistic--physical embodied agents are extremely large, which presents a serious challenge to classical einforcement learning schemes. In previous work--(Jacob et al., 2005a, Jacob et al., 2005b) we introduced--our EMBER (for EMbodiment-Based modulaR) reinforcement learning system, which describes a novel method for decomposing agents into modules based on the agent s embodiment. This modular decomposition factorises the statespace--and dramatically improves performance--in unknown and dynamic environments. However,--while there are great advantages to be gained from a factorised state-space, the question of dependencies cannot be ignored. We present a development of the work reported in (Jacob et al., 2004) which shows, in a simple example, how dependencies may be identified using a heuristic approach. Results show that the--system is able quickly to discover and act upon--dependencies, even where they are neither simple--nor deterministic.en
dc.language.isoeng
dc.relation.ispartofTAROS
dc.titleInferring dependencies in Embodiment-based modular reinforcement learningen
dc.contributor.institutionSchool of Computer Science
dc.contributor.institutionCentre for Computer Science and Informatics Research
dc.contributor.institutionDepartment of Computer Science
dc.contributor.institutionAdaptive Systems
dc.contributor.institutionSchool of Physics, Engineering & Computer Science
dc.contributor.institutionCentre for Future Societies Research
dc.contributor.institutionScience & Technology Research Institute
dc.description.statusPeer reviewed
rioxxterms.typeJournal Article/Review
herts.preservation.rarelyaccessedtrue


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record