Hybrid Autoregressive Transducer (hat)

Hybrid Autoregressive Transducer (hat)

This paper proposes and evaluates the hybrid autoregressive transducer (HAT) model, a time-synchronous encoderdecoder model that preserves the modularity of conventional automatic speech recognition systems. The HAT model provides a way to measure the quality of the internal language model that can...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Variani, Ehsan, Rybach, David, Allauzen, Cyril, Riley, Michael
Format: Artikel
Sprache:eng
Schlagworte:
Computer Science - Computation and Language
Computer Science - Learning
Computer Science - Sound
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title
container_volume
creator Variani, Ehsan
Rybach, David
Allauzen, Cyril
Riley, Michael
description This paper proposes and evaluates the hybrid autoregressive transducer (HAT) model, a time-synchronous encoderdecoder model that preserves the modularity of conventional automatic speech recognition systems. The HAT model provides a way to measure the quality of the internal language model that can be used to decide whether inference with an external language model is beneficial or not. This article also presents a finite context version of the HAT model that addresses the exposure bias problem and significantly simplifies the overall training and inference. We evaluate our proposed model on a large-scale voice search task. Our experiments show significant improvements in WER compared to the state-of-the-art approaches.
doi_str_mv 10.48550/arxiv.2003.07705
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2003_07705</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2003_07705</sourcerecordid><originalsourceid>FETCH-LOGICAL-a675-2357c3ca38426e78736b069ce82ee8e18d267cee0d9d2572ae71b84a9197d8f53</originalsourceid><addsrcrecordid>eNotzrsOgjAUgOEuDkZ9ACeZjA5gaWlPGQ3xlpi4sJNDe1QSbylK5O2N6PRvfz7GxjGPEqMUX6B_V00kOJcRB-Cqz6bbtvSVC5av593TyVNdVw0Fucdb7V6WfDA743M-ZL0jXmoa_Ttg-XqVZ9twf9jssuU-RA0qFFKBlRalSYQmMCB1yXVqyQgiQ7FxQoMl4i51QoFAgrg0CaZxCs4clRywyW_bQYuHr67o2-ILLjqw_AAdYTmn</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Hybrid Autoregressive Transducer (hat)</title><source>arXiv.org</source><creator>Variani, Ehsan ; Rybach, David ; Allauzen, Cyril ; Riley, Michael</creator><creatorcontrib>Variani, Ehsan ; Rybach, David ; Allauzen, Cyril ; Riley, Michael</creatorcontrib><description>This paper proposes and evaluates the hybrid autoregressive transducer (HAT) model, a time-synchronous encoderdecoder model that preserves the modularity of conventional automatic speech recognition systems. The HAT model provides a way to measure the quality of the internal language model that can be used to decide whether inference with an external language model is beneficial or not. This article also presents a finite context version of the HAT model that addresses the exposure bias problem and significantly simplifies the overall training and inference. We evaluate our proposed model on a large-scale voice search task. Our experiments show significant improvements in WER compared to the state-of-the-art approaches.</description><identifier>DOI: 10.48550/arxiv.2003.07705</identifier><language>eng</language><subject>Computer Science - Computation and Language ; Computer Science - Learning ; Computer Science - Sound</subject><creationdate>2020-03</creationdate><rights>http://arxiv.org/licenses/nonexclusive-distrib/1.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,780,885</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2003.07705$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2003.07705$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Variani, Ehsan</creatorcontrib><creatorcontrib>Rybach, David</creatorcontrib><creatorcontrib>Allauzen, Cyril</creatorcontrib><creatorcontrib>Riley, Michael</creatorcontrib><title>Hybrid Autoregressive Transducer (hat)</title><description>This paper proposes and evaluates the hybrid autoregressive transducer (HAT) model, a time-synchronous encoderdecoder model that preserves the modularity of conventional automatic speech recognition systems. The HAT model provides a way to measure the quality of the internal language model that can be used to decide whether inference with an external language model is beneficial or not. This article also presents a finite context version of the HAT model that addresses the exposure bias problem and significantly simplifies the overall training and inference. We evaluate our proposed model on a large-scale voice search task. Our experiments show significant improvements in WER compared to the state-of-the-art approaches.</description><subject>Computer Science - Computation and Language</subject><subject>Computer Science - Learning</subject><subject>Computer Science - Sound</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2020</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNotzrsOgjAUgOEuDkZ9ACeZjA5gaWlPGQ3xlpi4sJNDe1QSbylK5O2N6PRvfz7GxjGPEqMUX6B_V00kOJcRB-Cqz6bbtvSVC5av593TyVNdVw0Fucdb7V6WfDA743M-ZL0jXmoa_Ttg-XqVZ9twf9jssuU-RA0qFFKBlRalSYQmMCB1yXVqyQgiQ7FxQoMl4i51QoFAgrg0CaZxCs4clRywyW_bQYuHr67o2-ILLjqw_AAdYTmn</recordid><startdate>20200312</startdate><enddate>20200312</enddate><creator>Variani, Ehsan</creator><creator>Rybach, David</creator><creator>Allauzen, Cyril</creator><creator>Riley, Michael</creator><scope>AKY</scope><scope>GOX</scope></search><sort><creationdate>20200312</creationdate><title>Hybrid Autoregressive Transducer (hat)</title><author>Variani, Ehsan ; Rybach, David ; Allauzen, Cyril ; Riley, Michael</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a675-2357c3ca38426e78736b069ce82ee8e18d267cee0d9d2572ae71b84a9197d8f53</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2020</creationdate><topic>Computer Science - Computation and Language</topic><topic>Computer Science - Learning</topic><topic>Computer Science - Sound</topic><toplevel>online_resources</toplevel><creatorcontrib>Variani, Ehsan</creatorcontrib><creatorcontrib>Rybach, David</creatorcontrib><creatorcontrib>Allauzen, Cyril</creatorcontrib><creatorcontrib>Riley, Michael</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Variani, Ehsan</au><au>Rybach, David</au><au>Allauzen, Cyril</au><au>Riley, Michael</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Hybrid Autoregressive Transducer (hat)</atitle><date>2020-03-12</date><risdate>2020</risdate><abstract>This paper proposes and evaluates the hybrid autoregressive transducer (HAT) model, a time-synchronous encoderdecoder model that preserves the modularity of conventional automatic speech recognition systems. The HAT model provides a way to measure the quality of the internal language model that can be used to decide whether inference with an external language model is beneficial or not. This article also presents a finite context version of the HAT model that addresses the exposure bias problem and significantly simplifies the overall training and inference. We evaluate our proposed model on a large-scale voice search task. Our experiments show significant improvements in WER compared to the state-of-the-art approaches.</abstract><doi>10.48550/arxiv.2003.07705</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.2003.07705
ispartof
issn
language eng
recordid cdi_arxiv_primary_2003_07705
source arXiv.org
subjects Computer Science - Computation and Language
Computer Science - Learning
Computer Science - Sound
title Hybrid Autoregressive Transducer (hat)
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-27T14%3A24%3A57IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-arxiv_GOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Hybrid%20Autoregressive%20Transducer%20(hat)&rft.au=Variani,%20Ehsan&rft.date=2020-03-12&rft_id=info:doi/10.48550/arxiv.2003.07705&rft_dat=%3Carxiv_GOX%3E2003_07705%3C/arxiv_GOX%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true