Stick-breaking construction for the Indian buffet process

The Indian buffet process (IBP) is a Bayesian nonparametric distribution whereby objects are modelled using an unbounded number of latent features. In this paper we derive a stick-breaking representation for the IBP. Based on this new representation, we develop slice samplers for the IBP that are ef...

Full description

Bibliographic Details
Main Authors: Teh, Y, Görür, D, Ghahramani, Z
Format: Journal article
Language:English
Published: 2007
_version_ 1797074576292708352
author Teh, Y
Görür, D
Ghahramani, Z
author_facet Teh, Y
Görür, D
Ghahramani, Z
author_sort Teh, Y
collection OXFORD
description The Indian buffet process (IBP) is a Bayesian nonparametric distribution whereby objects are modelled using an unbounded number of latent features. In this paper we derive a stick-breaking representation for the IBP. Based on this new representation, we develop slice samplers for the IBP that are efficient, easy to implement and are more generally applicable than the currently available Gibbs sampler. This representation, along with the work of Thibaux and Jordan [17], also illuminates interesting theoretical connections between the IBP, Chinese restaurant processes, Beta processes and Dirichlet processes.
first_indexed 2024-03-06T23:38:13Z
format Journal article
id oxford-uuid:6e65f1e3-6181-4cb3-9474-bd4ae391cd0f
institution University of Oxford
language English
last_indexed 2024-03-06T23:38:13Z
publishDate 2007
record_format dspace
spelling oxford-uuid:6e65f1e3-6181-4cb3-9474-bd4ae391cd0f2022-03-26T19:24:10ZStick-breaking construction for the Indian buffet processJournal articlehttp://purl.org/coar/resource_type/c_dcae04bcuuid:6e65f1e3-6181-4cb3-9474-bd4ae391cd0fEnglishSymplectic Elements at Oxford2007Teh, YGörür, DGhahramani, ZThe Indian buffet process (IBP) is a Bayesian nonparametric distribution whereby objects are modelled using an unbounded number of latent features. In this paper we derive a stick-breaking representation for the IBP. Based on this new representation, we develop slice samplers for the IBP that are efficient, easy to implement and are more generally applicable than the currently available Gibbs sampler. This representation, along with the work of Thibaux and Jordan [17], also illuminates interesting theoretical connections between the IBP, Chinese restaurant processes, Beta processes and Dirichlet processes.
spellingShingle Teh, Y
Görür, D
Ghahramani, Z
Stick-breaking construction for the Indian buffet process
title Stick-breaking construction for the Indian buffet process
title_full Stick-breaking construction for the Indian buffet process
title_fullStr Stick-breaking construction for the Indian buffet process
title_full_unstemmed Stick-breaking construction for the Indian buffet process
title_short Stick-breaking construction for the Indian buffet process
title_sort stick breaking construction for the indian buffet process
work_keys_str_mv AT tehy stickbreakingconstructionfortheindianbuffetprocess
AT gorurd stickbreakingconstructionfortheindianbuffetprocess
AT ghahramaniz stickbreakingconstructionfortheindianbuffetprocess