The IBM 2011 GALE Arabic speech transcription system
We describe the Arabic broadcast transcription system fielded by IBM in the GALE Phase 5 machine translation evaluation. Key advances over our Phase 4 system include a new Bayesian Sensing HMM acoustic model; multistream neural network features; a MADA vowelized acoustic model; and the use of a vari...
Gespeichert in:
| Hauptverfasser: | , , , , , , |
|---|---|
| Format: | Tagungsbericht |
| Sprache: | eng |
| Schlagworte: | |
| Online-Zugang: | Volltext bestellen |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| container_end_page | 277 |
|---|---|
| container_issue | |
| container_start_page | 272 |
| container_title | |
| container_volume | |
| creator | Mangu, L. Hong-Kwang Kuo Chu, S. Kingsbury, B. Saon, G. Soltau, H. Biadsy, F. |
| description | We describe the Arabic broadcast transcription system fielded by IBM in the GALE Phase 5 machine translation evaluation. Key advances over our Phase 4 system include a new Bayesian Sensing HMM acoustic model; multistream neural network features; a MADA vowelized acoustic model; and the use of a variety of language model techniques with significant additive gains. These advances were instrumental in achieving a word error rate of 7.4% on the Phase 5 evaluation set, and an absolute improvement of 0.9% word error rate over our 2009 system on the unsequestered Phase 4 evaluation data. |
| doi_str_mv | 10.1109/ASRU.2011.6163943 |
| format | Conference Proceeding |
| fullrecord | <record><control><sourceid>ieee_6IE</sourceid><recordid>TN_cdi_ieee_primary_6163943</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><ieee_id>6163943</ieee_id><sourcerecordid>6163943</sourcerecordid><originalsourceid>FETCH-LOGICAL-i90t-1eaec1dbd8f3d43af6bba046479c08cdbbd01ab6169df489845eb1e01ac946bb3</originalsourceid><addsrcrecordid>eNo1T81Kw0AYXJFCteYBipd9gcT93M3-HGOptRARND2X_flCV2wNu7n07Y1Y5zLMMAwzhCyBVQDMPDQf77vqkQFUEiQ3gl-RwigNQirOuFT1Nbn9F7WekyLnTzZBSq2kuSGiOyDdPr3S3w66ado1bZJ10dM8IPoDHZM9ZZ_iMMbvE83nPOLxjsx6-5WxuPCCdM_rbvVStm-b7appy2jYWAJa9BBc0D0PgtteOmeZkEIZz7QPzgUG1k27TeiFNlrU6AAnzxsxZfmC3P_VRkTcDykebTrvLz_5D41URRE</addsrcrecordid><sourcetype>Publisher</sourcetype><iscdi>true</iscdi><recordtype>conference_proceeding</recordtype></control><display><type>conference_proceeding</type><title>The IBM 2011 GALE Arabic speech transcription system</title><source>IEEE Electronic Library (IEL) Conference Proceedings</source><creator>Mangu, L. ; Hong-Kwang Kuo ; Chu, S. ; Kingsbury, B. ; Saon, G. ; Soltau, H. ; Biadsy, F.</creator><creatorcontrib>Mangu, L. ; Hong-Kwang Kuo ; Chu, S. ; Kingsbury, B. ; Saon, G. ; Soltau, H. ; Biadsy, F.</creatorcontrib><description>We describe the Arabic broadcast transcription system fielded by IBM in the GALE Phase 5 machine translation evaluation. Key advances over our Phase 4 system include a new Bayesian Sensing HMM acoustic model; multistream neural network features; a MADA vowelized acoustic model; and the use of a variety of language model techniques with significant additive gains. These advances were instrumental in achieving a word error rate of 7.4% on the Phase 5 evaluation set, and an absolute improvement of 0.9% word error rate over our 2009 system on the unsequestered Phase 4 evaluation data.</description><identifier>ISBN: 1467303658</identifier><identifier>ISBN: 9781467303651</identifier><identifier>EISBN: 9781467303675</identifier><identifier>EISBN: 1467303666</identifier><identifier>EISBN: 9781467303668</identifier><identifier>EISBN: 1467303674</identifier><identifier>DOI: 10.1109/ASRU.2011.6163943</identifier><language>eng</language><publisher>IEEE</publisher><subject>Acoustics ; Computational modeling ; Dictionaries ; Hidden Markov models ; large vocabulary speech recognition ; Lattices ; Training ; Transforms</subject><ispartof>2011 IEEE Workshop on Automatic Speech Recognition & Understanding, 2011, p.272-277</ispartof><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://ieeexplore.ieee.org/document/6163943$$EHTML$$P50$$Gieee$$H</linktohtml><link.rule.ids>309,310,780,784,789,790,2058,27925,54920</link.rule.ids><linktorsrc>$$Uhttps://ieeexplore.ieee.org/document/6163943$$EView_record_in_IEEE$$FView_record_in_$$GIEEE</linktorsrc></links><search><creatorcontrib>Mangu, L.</creatorcontrib><creatorcontrib>Hong-Kwang Kuo</creatorcontrib><creatorcontrib>Chu, S.</creatorcontrib><creatorcontrib>Kingsbury, B.</creatorcontrib><creatorcontrib>Saon, G.</creatorcontrib><creatorcontrib>Soltau, H.</creatorcontrib><creatorcontrib>Biadsy, F.</creatorcontrib><title>The IBM 2011 GALE Arabic speech transcription system</title><title>2011 IEEE Workshop on Automatic Speech Recognition & Understanding</title><addtitle>ASRU</addtitle><description>We describe the Arabic broadcast transcription system fielded by IBM in the GALE Phase 5 machine translation evaluation. Key advances over our Phase 4 system include a new Bayesian Sensing HMM acoustic model; multistream neural network features; a MADA vowelized acoustic model; and the use of a variety of language model techniques with significant additive gains. These advances were instrumental in achieving a word error rate of 7.4% on the Phase 5 evaluation set, and an absolute improvement of 0.9% word error rate over our 2009 system on the unsequestered Phase 4 evaluation data.</description><subject>Acoustics</subject><subject>Computational modeling</subject><subject>Dictionaries</subject><subject>Hidden Markov models</subject><subject>large vocabulary speech recognition</subject><subject>Lattices</subject><subject>Training</subject><subject>Transforms</subject><isbn>1467303658</isbn><isbn>9781467303651</isbn><isbn>9781467303675</isbn><isbn>1467303666</isbn><isbn>9781467303668</isbn><isbn>1467303674</isbn><fulltext>true</fulltext><rsrctype>conference_proceeding</rsrctype><creationdate>2011</creationdate><recordtype>conference_proceeding</recordtype><sourceid>6IE</sourceid><sourceid>RIE</sourceid><recordid>eNo1T81Kw0AYXJFCteYBipd9gcT93M3-HGOptRARND2X_flCV2wNu7n07Y1Y5zLMMAwzhCyBVQDMPDQf77vqkQFUEiQ3gl-RwigNQirOuFT1Nbn9F7WekyLnTzZBSq2kuSGiOyDdPr3S3w66ado1bZJ10dM8IPoDHZM9ZZ_iMMbvE83nPOLxjsx6-5WxuPCCdM_rbvVStm-b7appy2jYWAJa9BBc0D0PgtteOmeZkEIZz7QPzgUG1k27TeiFNlrU6AAnzxsxZfmC3P_VRkTcDykebTrvLz_5D41URRE</recordid><startdate>201112</startdate><enddate>201112</enddate><creator>Mangu, L.</creator><creator>Hong-Kwang Kuo</creator><creator>Chu, S.</creator><creator>Kingsbury, B.</creator><creator>Saon, G.</creator><creator>Soltau, H.</creator><creator>Biadsy, F.</creator><general>IEEE</general><scope>6IE</scope><scope>6IL</scope><scope>CBEJK</scope><scope>RIE</scope><scope>RIL</scope></search><sort><creationdate>201112</creationdate><title>The IBM 2011 GALE Arabic speech transcription system</title><author>Mangu, L. ; Hong-Kwang Kuo ; Chu, S. ; Kingsbury, B. ; Saon, G. ; Soltau, H. ; Biadsy, F.</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-i90t-1eaec1dbd8f3d43af6bba046479c08cdbbd01ab6169df489845eb1e01ac946bb3</frbrgroupid><rsrctype>conference_proceedings</rsrctype><prefilter>conference_proceedings</prefilter><language>eng</language><creationdate>2011</creationdate><topic>Acoustics</topic><topic>Computational modeling</topic><topic>Dictionaries</topic><topic>Hidden Markov models</topic><topic>large vocabulary speech recognition</topic><topic>Lattices</topic><topic>Training</topic><topic>Transforms</topic><toplevel>online_resources</toplevel><creatorcontrib>Mangu, L.</creatorcontrib><creatorcontrib>Hong-Kwang Kuo</creatorcontrib><creatorcontrib>Chu, S.</creatorcontrib><creatorcontrib>Kingsbury, B.</creatorcontrib><creatorcontrib>Saon, G.</creatorcontrib><creatorcontrib>Soltau, H.</creatorcontrib><creatorcontrib>Biadsy, F.</creatorcontrib><collection>IEEE Electronic Library (IEL) Conference Proceedings</collection><collection>IEEE Proceedings Order Plan All Online (POP All Online) 1998-present by volume</collection><collection>IEEE Xplore All Conference Proceedings</collection><collection>IEEE Electronic Library (IEL)</collection><collection>IEEE Proceedings Order Plans (POP All) 1998-Present</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Mangu, L.</au><au>Hong-Kwang Kuo</au><au>Chu, S.</au><au>Kingsbury, B.</au><au>Saon, G.</au><au>Soltau, H.</au><au>Biadsy, F.</au><format>book</format><genre>proceeding</genre><ristype>CONF</ristype><atitle>The IBM 2011 GALE Arabic speech transcription system</atitle><btitle>2011 IEEE Workshop on Automatic Speech Recognition & Understanding</btitle><stitle>ASRU</stitle><date>2011-12</date><risdate>2011</risdate><spage>272</spage><epage>277</epage><pages>272-277</pages><isbn>1467303658</isbn><isbn>9781467303651</isbn><eisbn>9781467303675</eisbn><eisbn>1467303666</eisbn><eisbn>9781467303668</eisbn><eisbn>1467303674</eisbn><abstract>We describe the Arabic broadcast transcription system fielded by IBM in the GALE Phase 5 machine translation evaluation. Key advances over our Phase 4 system include a new Bayesian Sensing HMM acoustic model; multistream neural network features; a MADA vowelized acoustic model; and the use of a variety of language model techniques with significant additive gains. These advances were instrumental in achieving a word error rate of 7.4% on the Phase 5 evaluation set, and an absolute improvement of 0.9% word error rate over our 2009 system on the unsequestered Phase 4 evaluation data.</abstract><pub>IEEE</pub><doi>10.1109/ASRU.2011.6163943</doi><tpages>6</tpages></addata></record> |
| fulltext | fulltext_linktorsrc |
| identifier | ISBN: 1467303658 |
| ispartof | 2011 IEEE Workshop on Automatic Speech Recognition & Understanding, 2011, p.272-277 |
| issn | |
| language | eng |
| recordid | cdi_ieee_primary_6163943 |
| source | IEEE Electronic Library (IEL) Conference Proceedings |
| subjects | Acoustics Computational modeling Dictionaries Hidden Markov models large vocabulary speech recognition Lattices Training Transforms |
| title | The IBM 2011 GALE Arabic speech transcription system |
| url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-05T06%3A31%3A53IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-ieee_6IE&rft_val_fmt=info:ofi/fmt:kev:mtx:book&rft.genre=proceeding&rft.atitle=The%20IBM%202011%20GALE%20Arabic%20speech%20transcription%20system&rft.btitle=2011%20IEEE%20Workshop%20on%20Automatic%20Speech%20Recognition%20&%20Understanding&rft.au=Mangu,%20L.&rft.date=2011-12&rft.spage=272&rft.epage=277&rft.pages=272-277&rft.isbn=1467303658&rft.isbn_list=9781467303651&rft_id=info:doi/10.1109/ASRU.2011.6163943&rft_dat=%3Cieee_6IE%3E6163943%3C/ieee_6IE%3E%3Curl%3E%3C/url%3E&rft.eisbn=9781467303675&rft.eisbn_list=1467303666&rft.eisbn_list=9781467303668&rft.eisbn_list=1467303674&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rft_ieee_id=6163943&rfr_iscdi=true |