File Download
Supplementary

Conference Paper: Dependency parsing with energy-based reinforcement learning

TitleDependency parsing with energy-based reinforcement learning
Authors
Issue Date2009
Citation
The 11th International Conference on Parsing Technologies (IWPT'09), Paris, France, 7-9 October 2009. In Proceedings of the 11th IWPT, 2009, p. 234-237 How to Cite?
AbstractWe present a model which integrates dependency parsing with reinforcement learning based on Markov decision process. At each time step, a transition is picked up to construct the dependency tree in terms of the long-run reward. The optimal policy for choosing transitions can be found with the SARSA algorithm. In SARSA, an approximation of the state-action function can be obtained by calculating the negative free energies for the Restricted Boltzmann Machine. The experimental results on CoNLL-X multilingual data show that the proposed model achieves comparable results with the current state-of-the-art methods.
DescriptionShort Paper Session VII
Persistent Identifierhttp://hdl.handle.net/10722/142602

 

DC FieldValueLanguage
dc.contributor.authorZhang, Len_US
dc.contributor.authorChan, KPen_US
dc.date.accessioned2011-10-28T02:52:51Z-
dc.date.available2011-10-28T02:52:51Z-
dc.date.issued2009en_US
dc.identifier.citationThe 11th International Conference on Parsing Technologies (IWPT'09), Paris, France, 7-9 October 2009. In Proceedings of the 11th IWPT, 2009, p. 234-237en_US
dc.identifier.urihttp://hdl.handle.net/10722/142602-
dc.descriptionShort Paper Session VII-
dc.description.abstractWe present a model which integrates dependency parsing with reinforcement learning based on Markov decision process. At each time step, a transition is picked up to construct the dependency tree in terms of the long-run reward. The optimal policy for choosing transitions can be found with the SARSA algorithm. In SARSA, an approximation of the state-action function can be obtained by calculating the negative free energies for the Restricted Boltzmann Machine. The experimental results on CoNLL-X multilingual data show that the proposed model achieves comparable results with the current state-of-the-art methods.-
dc.languageengen_US
dc.relation.ispartofProceedings of the 11th International Conference on Parsing Technology, IWPT'09en_US
dc.rightsCreative Commons: Attribution 3.0 Hong Kong License-
dc.titleDependency parsing with energy-based reinforcement learningen_US
dc.typeConference_Paperen_US
dc.identifier.emailZhang, L: lzhang@cs.hku.hken_US
dc.identifier.emailChan, KP: kpchan@cs.hku.hk-
dc.identifier.authorityChan, KP=rp00092en_US
dc.description.naturepostprint-
dc.identifier.hkuros184437en_US
dc.identifier.spage234-
dc.identifier.epage237-
dc.description.otherThe 11th International Conference on Parsing Technologies (IWPT'09), Paris, France, 7-9 October 2009. In Proceedings of the 11th IWPT, 2009, p. 234-237-

Export via OAI-PMH Interface in XML Formats


OR


Export to Other Non-XML Formats