Using generation for grammar analysis and error detection
We demonstrate that the bidirectionality of deep grammars, allowing them to generate as well as parse sentences, can be used to automatically and effectively identify errors in the grammars. The system is tested on two implemented HPSG grammars: Jacy for Japanese, and the ERG for English. Using this...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Conference Paper |
| Language: | English |
| Published: |
2009
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/93757 http://hdl.handle.net/10220/6158 http://www.acl-ijcnlp-2009.org/main/acceptedshortpapers.html |
| _version_ | 1824454022052970496 |
|---|---|
| author | Goodman, Michael Wayne Bond, Francis |
| author2 | School of Humanities and Social Sciences |
| author_facet | School of Humanities and Social Sciences Goodman, Michael Wayne Bond, Francis |
| author_sort | Goodman, Michael Wayne |
| collection | NTU |
| description | We demonstrate that the bidirectionality of deep grammars, allowing them to generate as well as parse sentences, can be used to automatically and effectively identify errors in the grammars. The system is tested on two implemented HPSG grammars: Jacy for Japanese, and the ERG for English. Using this system, we were able to increase generation coverage in Jacy by 18% (45% to 63%) with only four weeks of grammar development. |
| first_indexed | 2025-02-19T03:15:42Z |
| format | Conference Paper |
| id | ntu-10356/93757 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2025-02-19T03:15:42Z |
| publishDate | 2009 |
| record_format | dspace |
| spelling | ntu-10356/937572019-12-06T18:44:58Z Using generation for grammar analysis and error detection Goodman, Michael Wayne Bond, Francis School of Humanities and Social Sciences Annual Meeting of the Association for Computational Linguistics (47th : 2009 : Singapore) DRNTU::Humanities::Linguistics We demonstrate that the bidirectionality of deep grammars, allowing them to generate as well as parse sentences, can be used to automatically and effectively identify errors in the grammars. The system is tested on two implemented HPSG grammars: Jacy for Japanese, and the ERG for English. Using this system, we were able to increase generation coverage in Jacy by 18% (45% to 63%) with only four weeks of grammar development. Accepted version 2009-11-13T06:51:47Z 2019-12-06T18:44:58Z 2009-11-13T06:51:47Z 2019-12-06T18:44:58Z 2009 2009 Conference Paper Goodman, M. W., & Bond, F. (2009). Using generation for grammar analysis and error detection. Proceedings of the ACL-IJCNLP 2009 Conference Short Papers (2009:Singapore): pp. 109–112. https://hdl.handle.net/10356/93757 http://hdl.handle.net/10220/6158 http://www.acl-ijcnlp-2009.org/main/acceptedshortpapers.html 148255 en Proceedings of the ACL-IJCNLP 2009 Conference Short Papers (2009:Singapore) © copyright 2009 ACL and AFNLP. The conference paper is available at www.acl-ijcnlp-2009.org/ 5 p. application/pdf |
| spellingShingle | DRNTU::Humanities::Linguistics Goodman, Michael Wayne Bond, Francis Using generation for grammar analysis and error detection |
| title | Using generation for grammar analysis and error detection |
| title_full | Using generation for grammar analysis and error detection |
| title_fullStr | Using generation for grammar analysis and error detection |
| title_full_unstemmed | Using generation for grammar analysis and error detection |
| title_short | Using generation for grammar analysis and error detection |
| title_sort | using generation for grammar analysis and error detection |
| topic | DRNTU::Humanities::Linguistics |
| url | https://hdl.handle.net/10356/93757 http://hdl.handle.net/10220/6158 http://www.acl-ijcnlp-2009.org/main/acceptedshortpapers.html |
| work_keys_str_mv | AT goodmanmichaelwayne usinggenerationforgrammaranalysisanderrordetection AT bondfrancis usinggenerationforgrammaranalysisanderrordetection |