Show simple item record

dc.contributor.authorTruyen, Tran
dc.contributor.authorPhung, Dinh
dc.contributor.authorVenkatesh, Svetha
dc.contributor.authorBui, Hung H.
dc.contributor.editorLi Deng, Dong Yu and Geoff Hinton
dc.identifier.citationTruyen, Tran The and Phung, Dinh Q. and Venkatesh, Svetha and Bui, Hung H. 2009. MCMC for Hierarchical Semi-Markov Conditional Random fields, in Li Deng, Dong Yu and Geoff Hinton (ed), NIPS 2009, Dec 12 2009, pp. 1-8. Whistler, BC, Canada.

Deep architecture such as hierarchical semi-Markov models is an important class of models for nested sequential data. Current exact inference schemes either cost cubic time in sequence length, or exponential time in model depth. These costs are prohibitive for large-scale problems with arbitrary length and depth. In this contribution, we propose a new approximation technique that may have the potential to achieve sub-cubic time complexity in length and linear time depth, at the cost of some loss of quality. The idea is based on two well-known methods: Gibbs sampling and Rao-Blackwellisation. We provide some simulation-based evaluation of the quality of the RGBS with respect to run time and sequence length.

dc.titleMCMC for Hierarchical Semi-Markov Conditional Random fields
dc.typeConference Paper
dcterms.source.titleDeep learning for speech recognition and related applications
dcterms.source.seriesDeep learning for speech recognition and related applications
dcterms.source.conferenceNIPS 2009
dcterms.source.conference-start-dateDec 12 2009
dcterms.source.conferencelocationWhister,BC, Canada
dcterms.source.placeWhistler, Canada
curtin.accessStatusOpen access
curtin.facultySchool of Science and Computing
curtin.facultyDepartment of Computing
curtin.facultyFaculty of Science and Engineering

Files in this item


This item appears in the following Collection(s)

Show simple item record