Bayesian Synchronous Grammar Induction.
We present a novel method for inducing synchronous context free grammars (SCFGs) from a corpus of parallel string pairs. SCFGs can model equivalence between strings in terms of substitutions, insertions and deletions, and the reordering of sub-strings. We develop a non-parametric Bayesian model and...
| Main Authors: | , , |
|---|---|
| Other Authors: | |
| Format: | Conference item |
| Published: |
Curran Associates, Inc.
2008
|
| _version_ | 1797101741867532288 |
|---|---|
| author | Blunsom, P Cohn, T Osborne, M |
| author2 | Koller, D |
| author_facet | Koller, D Blunsom, P Cohn, T Osborne, M |
| author_sort | Blunsom, P |
| collection | OXFORD |
| description | We present a novel method for inducing synchronous context free grammars (SCFGs) from a corpus of parallel string pairs. SCFGs can model equivalence between strings in terms of substitutions, insertions and deletions, and the reordering of sub-strings. We develop a non-parametric Bayesian model and apply it to a machine translation task, using priors to replace the various heuristics commonly used in this field. Using a variational Bayes training procedure, we learn the latent structure of translation equivalence through the induction of synchronous grammar categories for phrasal translations, showing improvements in translation performance over maximum likelihood models. |
| first_indexed | 2024-03-07T05:56:10Z |
| format | Conference item |
| id | oxford-uuid:ea936cae-8be4-400a-8f12-8ff88b877cc2 |
| institution | University of Oxford |
| last_indexed | 2024-03-07T05:56:10Z |
| publishDate | 2008 |
| publisher | Curran Associates, Inc. |
| record_format | dspace |
| spelling | oxford-uuid:ea936cae-8be4-400a-8f12-8ff88b877cc22022-03-27T11:03:16ZBayesian Synchronous Grammar Induction.Conference itemhttp://purl.org/coar/resource_type/c_5794uuid:ea936cae-8be4-400a-8f12-8ff88b877cc2Symplectic Elements at OxfordCurran Associates, Inc.2008Blunsom, PCohn, TOsborne, MKoller, DSchuurmans, DBengio, YBottou, LWe present a novel method for inducing synchronous context free grammars (SCFGs) from a corpus of parallel string pairs. SCFGs can model equivalence between strings in terms of substitutions, insertions and deletions, and the reordering of sub-strings. We develop a non-parametric Bayesian model and apply it to a machine translation task, using priors to replace the various heuristics commonly used in this field. Using a variational Bayes training procedure, we learn the latent structure of translation equivalence through the induction of synchronous grammar categories for phrasal translations, showing improvements in translation performance over maximum likelihood models. |
| spellingShingle | Blunsom, P Cohn, T Osborne, M Bayesian Synchronous Grammar Induction. |
| title | Bayesian Synchronous Grammar Induction. |
| title_full | Bayesian Synchronous Grammar Induction. |
| title_fullStr | Bayesian Synchronous Grammar Induction. |
| title_full_unstemmed | Bayesian Synchronous Grammar Induction. |
| title_short | Bayesian Synchronous Grammar Induction. |
| title_sort | bayesian synchronous grammar induction |
| work_keys_str_mv | AT blunsomp bayesiansynchronousgrammarinduction AT cohnt bayesiansynchronousgrammarinduction AT osbornem bayesiansynchronousgrammarinduction |