LAST: Scalable Lattice-Based Speech Modelling in JAX

We introduce LAST, a LAttice-based Speech Transducer library in JAX. With an emphasis on flexibility, ease-of-use, and scalability, LAST implements differentiable weighted finite state automaton (WFSA) algorithms needed for training \& inference that scale to a large WFSA such as a recognition l...

Full description

Saved in:
Bibliographic Details
Published in:arXiv.org 2023-04
Main Authors: Wu, Ke, Variani, Ehsan, Bagby, Tom, Riley, Michael
Format: Article
Language:English
Subjects:
Algorithms
Speech
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
cited_by
cites
container_end_page
container_issue
container_start_page
container_title arXiv.org
container_volume
creator Wu, Ke
Variani, Ehsan
Bagby, Tom
Riley, Michael
description We introduce LAST, a LAttice-based Speech Transducer library in JAX. With an emphasis on flexibility, ease-of-use, and scalability, LAST implements differentiable weighted finite state automaton (WFSA) algorithms needed for training \& inference that scale to a large WFSA such as a recognition lattice over the entire utterance. Despite these WFSA algorithms being well-known in the literature, new challenges arise from performance characteristics of modern architectures, and from nuances in automatic differentiation. We describe a suite of generally applicable techniques employed in LAST to address these challenges, and demonstrate their effectiveness with benchmarks on TPUv3 and V100 GPU.
format article
fullrecord <record><control><sourceid>proquest</sourceid><recordid>TN_cdi_proquest_journals_2806661079</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2806661079</sourcerecordid><originalsourceid>FETCH-proquest_journals_28066610793</originalsourceid><addsrcrecordid>eNpjYuA0MjY21LUwMTLiYOAtLs4yMDAwMjM3MjU15mQw8XEMDrFSCE5OzElMyklV8EksKclMTtV1SixOTVEILkhNTc5Q8M1PSc3JycxLV8jMU_ByjOBhYE1LzClO5YXS3AzKbq4hzh66BUX5haWpxSXxWfmlRXlAqXgjCwMzMzNDA3NLY-JUAQCARzLX</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2806661079</pqid></control><display><type>article</type><title>LAST: Scalable Lattice-Based Speech Modelling in JAX</title><source>Publicly Available Content Database</source><creator>Wu, Ke ; Variani, Ehsan ; Bagby, Tom ; Riley, Michael</creator><creatorcontrib>Wu, Ke ; Variani, Ehsan ; Bagby, Tom ; Riley, Michael</creatorcontrib><description>We introduce LAST, a LAttice-based Speech Transducer library in JAX. With an emphasis on flexibility, ease-of-use, and scalability, LAST implements differentiable weighted finite state automaton (WFSA) algorithms needed for training \&amp; inference that scale to a large WFSA such as a recognition lattice over the entire utterance. Despite these WFSA algorithms being well-known in the literature, new challenges arise from performance characteristics of modern architectures, and from nuances in automatic differentiation. We describe a suite of generally applicable techniques employed in LAST to address these challenges, and demonstrate their effectiveness with benchmarks on TPUv3 and V100 GPU.</description><identifier>EISSN: 2331-8422</identifier><language>eng</language><publisher>Ithaca: Cornell University Library, arXiv.org</publisher><subject>Algorithms ; Speech</subject><ispartof>arXiv.org, 2023-04</ispartof><rights>2023. This work is published under http://creativecommons.org/licenses/by/4.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://www.proquest.com/docview/2806661079?pq-origsite=primo$$EHTML$$P50$$Gproquest$$Hfree_for_read</linktohtml><link.rule.ids>780,784,25752,37011,44589</link.rule.ids></links><search><creatorcontrib>Wu, Ke</creatorcontrib><creatorcontrib>Variani, Ehsan</creatorcontrib><creatorcontrib>Bagby, Tom</creatorcontrib><creatorcontrib>Riley, Michael</creatorcontrib><title>LAST: Scalable Lattice-Based Speech Modelling in JAX</title><title>arXiv.org</title><description>We introduce LAST, a LAttice-based Speech Transducer library in JAX. With an emphasis on flexibility, ease-of-use, and scalability, LAST implements differentiable weighted finite state automaton (WFSA) algorithms needed for training \&amp; inference that scale to a large WFSA such as a recognition lattice over the entire utterance. Despite these WFSA algorithms being well-known in the literature, new challenges arise from performance characteristics of modern architectures, and from nuances in automatic differentiation. We describe a suite of generally applicable techniques employed in LAST to address these challenges, and demonstrate their effectiveness with benchmarks on TPUv3 and V100 GPU.</description><subject>Algorithms</subject><subject>Speech</subject><issn>2331-8422</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2023</creationdate><recordtype>article</recordtype><sourceid>PIMPY</sourceid><recordid>eNpjYuA0MjY21LUwMTLiYOAtLs4yMDAwMjM3MjU15mQw8XEMDrFSCE5OzElMyklV8EksKclMTtV1SixOTVEILkhNTc5Q8M1PSc3JycxLV8jMU_ByjOBhYE1LzClO5YXS3AzKbq4hzh66BUX5haWpxSXxWfmlRXlAqXgjCwMzMzNDA3NLY-JUAQCARzLX</recordid><startdate>20230425</startdate><enddate>20230425</enddate><creator>Wu, Ke</creator><creator>Variani, Ehsan</creator><creator>Bagby, Tom</creator><creator>Riley, Michael</creator><general>Cornell University Library, arXiv.org</general><scope>8FE</scope><scope>8FG</scope><scope>ABJCF</scope><scope>ABUWG</scope><scope>AFKRA</scope><scope>AZQEC</scope><scope>BENPR</scope><scope>BGLVJ</scope><scope>CCPQU</scope><scope>DWQXO</scope><scope>HCIFZ</scope><scope>L6V</scope><scope>M7S</scope><scope>PIMPY</scope><scope>PQEST</scope><scope>PQQKQ</scope><scope>PQUKI</scope><scope>PRINS</scope><scope>PTHSS</scope></search><sort><creationdate>20230425</creationdate><title>LAST: Scalable Lattice-Based Speech Modelling in JAX</title><author>Wu, Ke ; Variani, Ehsan ; Bagby, Tom ; Riley, Michael</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-proquest_journals_28066610793</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2023</creationdate><topic>Algorithms</topic><topic>Speech</topic><toplevel>online_resources</toplevel><creatorcontrib>Wu, Ke</creatorcontrib><creatorcontrib>Variani, Ehsan</creatorcontrib><creatorcontrib>Bagby, Tom</creatorcontrib><creatorcontrib>Riley, Michael</creatorcontrib><collection>ProQuest SciTech Collection</collection><collection>ProQuest Technology Collection</collection><collection>Materials Science &amp; Engineering Collection</collection><collection>ProQuest Central (Alumni)</collection><collection>ProQuest Central</collection><collection>ProQuest Central Essentials</collection><collection>AUTh Library subscriptions: ProQuest Central</collection><collection>Technology Collection</collection><collection>ProQuest One Community College</collection><collection>ProQuest Central</collection><collection>SciTech Premium Collection</collection><collection>ProQuest Engineering Collection</collection><collection>Engineering Database</collection><collection>Publicly Available Content Database</collection><collection>ProQuest One Academic Eastern Edition (DO NOT USE)</collection><collection>ProQuest One Academic</collection><collection>ProQuest One Academic UKI Edition</collection><collection>ProQuest Central China</collection><collection>Engineering Collection</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Wu, Ke</au><au>Variani, Ehsan</au><au>Bagby, Tom</au><au>Riley, Michael</au><format>book</format><genre>document</genre><ristype>GEN</ristype><atitle>LAST: Scalable Lattice-Based Speech Modelling in JAX</atitle><jtitle>arXiv.org</jtitle><date>2023-04-25</date><risdate>2023</risdate><eissn>2331-8422</eissn><abstract>We introduce LAST, a LAttice-based Speech Transducer library in JAX. With an emphasis on flexibility, ease-of-use, and scalability, LAST implements differentiable weighted finite state automaton (WFSA) algorithms needed for training \&amp; inference that scale to a large WFSA such as a recognition lattice over the entire utterance. Despite these WFSA algorithms being well-known in the literature, new challenges arise from performance characteristics of modern architectures, and from nuances in automatic differentiation. We describe a suite of generally applicable techniques employed in LAST to address these challenges, and demonstrate their effectiveness with benchmarks on TPUv3 and V100 GPU.</abstract><cop>Ithaca</cop><pub>Cornell University Library, arXiv.org</pub><oa>free_for_read</oa></addata></record>
fulltext fulltext
identifier EISSN: 2331-8422
ispartof arXiv.org, 2023-04
issn 2331-8422
language eng
recordid cdi_proquest_journals_2806661079
source Publicly Available Content Database
subjects Algorithms
Speech
title LAST: Scalable Lattice-Based Speech Modelling in JAX
url http://sfxeu10.hosted.exlibrisgroup.com/loughborough?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-11T22%3A48%3A49IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest&rft_val_fmt=info:ofi/fmt:kev:mtx:book&rft.genre=document&rft.atitle=LAST:%20Scalable%20Lattice-Based%20Speech%20Modelling%20in%20JAX&rft.jtitle=arXiv.org&rft.au=Wu,%20Ke&rft.date=2023-04-25&rft.eissn=2331-8422&rft_id=info:doi/&rft_dat=%3Cproquest%3E2806661079%3C/proquest%3E%3Cgrp_id%3Ecdi_FETCH-proquest_journals_28066610793%3C/grp_id%3E%3Coa%3E%3C/oa%3E%3Curl%3E%3C/url%3E&rft_id=info:oai/&rft_pqid=2806661079&rft_id=info:pmid/&rfr_iscdi=true