A hierarchical nonparametric Bayesian approach to statistical language model domain adaptation
In this paper we present a doubly hierarchical Pitman-Yor process language model. Its bottom layer of hierarchy consists of multiple hierarchical Pitman-Yor process language models, one each for some number of domains. The novel top layer of hierarchy consists of a mechanism to couple together multi...
| Main Authors: | , |
|---|---|
| Format: | Journal article |
| Language: | English |
| Published: |
2009
|
| _version_ | 1826273004296863744 |
|---|---|
| author | Wood, F Teh, Y |
| author_facet | Wood, F Teh, Y |
| author_sort | Wood, F |
| collection | OXFORD |
| description | In this paper we present a doubly hierarchical Pitman-Yor process language model. Its bottom layer of hierarchy consists of multiple hierarchical Pitman-Yor process language models, one each for some number of domains. The novel top layer of hierarchy consists of a mechanism to couple together multiple language models such that they share statistical strength. Intuitively this sharing results in the "adaptation" of a latent shared language model to each domain. We introduce a general formalism capable of describing the overallmodel which we call the graphical Pitman-Yor process and explain how to perform Bayesian inference in it. We present encouraging language model domain adaptation results that both illustrate the potential benefits of our new model and suggest new avenues of inquiry. © 2009 by the authors. |
| first_indexed | 2024-03-06T22:21:32Z |
| format | Journal article |
| id | oxford-uuid:55389ff4-8b5b-408c-a4bf-267319c665c7 |
| institution | University of Oxford |
| language | English |
| last_indexed | 2024-03-06T22:21:32Z |
| publishDate | 2009 |
| record_format | dspace |
| spelling | oxford-uuid:55389ff4-8b5b-408c-a4bf-267319c665c72022-03-26T16:42:42ZA hierarchical nonparametric Bayesian approach to statistical language model domain adaptationJournal articlehttp://purl.org/coar/resource_type/c_dcae04bcuuid:55389ff4-8b5b-408c-a4bf-267319c665c7EnglishSymplectic Elements at Oxford2009Wood, FTeh, YIn this paper we present a doubly hierarchical Pitman-Yor process language model. Its bottom layer of hierarchy consists of multiple hierarchical Pitman-Yor process language models, one each for some number of domains. The novel top layer of hierarchy consists of a mechanism to couple together multiple language models such that they share statistical strength. Intuitively this sharing results in the "adaptation" of a latent shared language model to each domain. We introduce a general formalism capable of describing the overallmodel which we call the graphical Pitman-Yor process and explain how to perform Bayesian inference in it. We present encouraging language model domain adaptation results that both illustrate the potential benefits of our new model and suggest new avenues of inquiry. © 2009 by the authors. |
| spellingShingle | Wood, F Teh, Y A hierarchical nonparametric Bayesian approach to statistical language model domain adaptation |
| title | A hierarchical nonparametric Bayesian approach to statistical language model domain adaptation |
| title_full | A hierarchical nonparametric Bayesian approach to statistical language model domain adaptation |
| title_fullStr | A hierarchical nonparametric Bayesian approach to statistical language model domain adaptation |
| title_full_unstemmed | A hierarchical nonparametric Bayesian approach to statistical language model domain adaptation |
| title_short | A hierarchical nonparametric Bayesian approach to statistical language model domain adaptation |
| title_sort | hierarchical nonparametric bayesian approach to statistical language model domain adaptation |
| work_keys_str_mv | AT woodf ahierarchicalnonparametricbayesianapproachtostatisticallanguagemodeldomainadaptation AT tehy ahierarchicalnonparametricbayesianapproachtostatisticallanguagemodeldomainadaptation AT woodf hierarchicalnonparametricbayesianapproachtostatisticallanguagemodeldomainadaptation AT tehy hierarchicalnonparametricbayesianapproachtostatisticallanguagemodeldomainadaptation |