Discriminative training of hierarchical acoustic models for large vocabulary continuous speech recognition
In this paper we propose discriminative training of hierarchical acoustic models for large vocabulary continuous speech recognition tasks. After presenting our hierarchical modeling framework, we describe how the models can be generated with either minimum classification error or large-margin traini...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Language: | en_US |
| Published: |
Institute of Electrical and Electronics Engineers
2011
|
| Online Access: | http://hdl.handle.net/1721.1/60562 https://orcid.org/0000-0002-3097-360X |
| _version_ | 1811086937172213760 |
|---|---|
| author | Chang, Hung-An Glass, James R. |
| author2 | Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory |
| author_facet | Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory Chang, Hung-An Glass, James R. |
| author_sort | Chang, Hung-An |
| collection | MIT |
| description | In this paper we propose discriminative training of hierarchical acoustic models for large vocabulary continuous speech recognition tasks. After presenting our hierarchical modeling framework, we describe how the models can be generated with either minimum classification error or large-margin training. Experiments on a large vocabulary lecture transcription task show that the hierarchical model can yield more than 1.0% absolute word error rate reduction over non-hierarchical models for both kinds of discriminative training. |
| first_indexed | 2024-09-23T13:37:11Z |
| format | Article |
| id | mit-1721.1/60562 |
| institution | Massachusetts Institute of Technology |
| language | en_US |
| last_indexed | 2024-09-23T13:37:11Z |
| publishDate | 2011 |
| publisher | Institute of Electrical and Electronics Engineers |
| record_format | dspace |
| spelling | mit-1721.1/605622022-09-28T15:04:19Z Discriminative training of hierarchical acoustic models for large vocabulary continuous speech recognition Chang, Hung-An Glass, James R. Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science Glass, James R. Chang, Hung-An Glass, James R. In this paper we propose discriminative training of hierarchical acoustic models for large vocabulary continuous speech recognition tasks. After presenting our hierarchical modeling framework, we describe how the models can be generated with either minimum classification error or large-margin training. Experiments on a large vocabulary lecture transcription task show that the hierarchical model can yield more than 1.0% absolute word error rate reduction over non-hierarchical models for both kinds of discriminative training. Taiwan Merit Scholarship (Number NSC-095- SAF-I-564-040-TMS) 2011-01-14T13:39:59Z 2011-01-14T13:39:59Z 2009-05 2009-04 Article http://purl.org/eprint/type/ConferencePaper 978-1-4244-2353-8 1520-6149 INSPEC Accession Number: 10701095 http://hdl.handle.net/1721.1/60562 Hung-An Chang, and J.R. Glass. “Discriminative training of hierarchical acoustic models for large vocabulary continuous speech recognition.” Acoustics, Speech and Signal Processing, 2009. ICASSP 2009. IEEE International Conference on. 2009. 4481-4484. © Copyright 2009 IEEE https://orcid.org/0000-0002-3097-360X en_US http://dx.doi.org/10.1109/ICASSP.2009.4960625 IEEE International Conference on Acoustics, Speech and Signal Processing, 2009. ICASSP 2009. Article is made available in accordance with the publisher's policy and may be subject to US copyright law. Please refer to the publisher's site for terms of use. application/pdf Institute of Electrical and Electronics Engineers MIT web domain |
| spellingShingle | Chang, Hung-An Glass, James R. Discriminative training of hierarchical acoustic models for large vocabulary continuous speech recognition |
| title | Discriminative training of hierarchical acoustic models for large vocabulary continuous speech recognition |
| title_full | Discriminative training of hierarchical acoustic models for large vocabulary continuous speech recognition |
| title_fullStr | Discriminative training of hierarchical acoustic models for large vocabulary continuous speech recognition |
| title_full_unstemmed | Discriminative training of hierarchical acoustic models for large vocabulary continuous speech recognition |
| title_short | Discriminative training of hierarchical acoustic models for large vocabulary continuous speech recognition |
| title_sort | discriminative training of hierarchical acoustic models for large vocabulary continuous speech recognition |
| url | http://hdl.handle.net/1721.1/60562 https://orcid.org/0000-0002-3097-360X |
| work_keys_str_mv | AT changhungan discriminativetrainingofhierarchicalacousticmodelsforlargevocabularycontinuousspeechrecognition AT glassjamesr discriminativetrainingofhierarchicalacousticmodelsforlargevocabularycontinuousspeechrecognition |