{"id":"https://openalex.org/W4384519051","doi":"https://doi.org/10.1109/taslp.2023.3296019","title":"LegoNN: Building Modular Encoder-Decoder Models","display_name":"LegoNN: Building Modular Encoder-Decoder Models","publication_year":2023,"publication_date":"2023-01-01","ids":{"openalex":"https://openalex.org/W4384519051","doi":"https://doi.org/10.1109/taslp.2023.3296019"},"language":"en","primary_location":{"id":"doi:10.1109/taslp.2023.3296019","is_oa":false,"landing_page_url":"https://doi.org/10.1109/taslp.2023.3296019","pdf_url":null,"source":{"id":"https://openalex.org/S4210169297","display_name":"IEEE/ACM Transactions on Audio Speech and Language Processing","issn_l":"2329-9290","issn":["2329-9290","2329-9304"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE/ACM Transactions on Audio, Speech, and Language Processing","raw_type":"journal-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":false,"oa_status":"closed","oa_url":null,"any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5068873086","display_name":"Siddharth Dalmia","orcid":"https://orcid.org/0000-0003-0437-5988"},"institutions":[{"id":"https://openalex.org/I74973139","display_name":"Carnegie Mellon University","ror":"https://ror.org/05x2bcf33","country_code":"US","type":"education","lineage":["https://openalex.org/I74973139"]}],"countries":["US"],"is_corresponding":true,"raw_author_name":"Siddharth Dalmia","raw_affiliation_strings":["Carnegie Mellon University, Pittsburgh, PA, USA","Meta AI, USA"],"affiliations":[{"raw_affiliation_string":"Carnegie Mellon University, Pittsburgh, PA, USA","institution_ids":["https://openalex.org/I74973139"]},{"raw_affiliation_string":"Meta AI, USA","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5022671092","display_name":"Dmytro Okhonko","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Dmytro Okhonko","raw_affiliation_strings":["Samaya AI., Mountain View, CA, USA","Meta AI, USA"],"affiliations":[{"raw_affiliation_string":"Samaya AI., Mountain View, CA, USA","institution_ids":[]},{"raw_affiliation_string":"Meta AI, USA","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5057420103","display_name":"Michael Lewis","orcid":"https://orcid.org/0000-0002-7620-1891"},"institutions":[{"id":"https://openalex.org/I4210114444","display_name":"Meta (United States)","ror":"https://ror.org/01zbnvs85","country_code":"US","type":"company","lineage":["https://openalex.org/I4210114444"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Mike Lewis","raw_affiliation_strings":["Meta Platforms Inc., Menlo Park, CA, USA"],"affiliations":[{"raw_affiliation_string":"Meta Platforms Inc., Menlo Park, CA, USA","institution_ids":["https://openalex.org/I4210114444"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5016113002","display_name":"Sergey Edunov","orcid":null},"institutions":[{"id":"https://openalex.org/I4210114444","display_name":"Meta (United States)","ror":"https://ror.org/01zbnvs85","country_code":"US","type":"company","lineage":["https://openalex.org/I4210114444"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Sergey Edunov","raw_affiliation_strings":["Meta Platforms Inc., Menlo Park, CA, USA"],"affiliations":[{"raw_affiliation_string":"Meta Platforms Inc., Menlo Park, CA, USA","institution_ids":["https://openalex.org/I4210114444"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5001291873","display_name":"Shinji Watanabe","orcid":"https://orcid.org/0000-0002-5970-8631"},"institutions":[{"id":"https://openalex.org/I74973139","display_name":"Carnegie Mellon University","ror":"https://ror.org/05x2bcf33","country_code":"US","type":"education","lineage":["https://openalex.org/I74973139"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Shinji Watanabe","raw_affiliation_strings":["Carnegie Mellon University, Pittsburgh, PA, USA","Meta AI, USA"],"affiliations":[{"raw_affiliation_string":"Carnegie Mellon University, Pittsburgh, PA, USA","institution_ids":["https://openalex.org/I74973139"]},{"raw_affiliation_string":"Meta AI, USA","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5085262529","display_name":"Florian Metze","orcid":"https://orcid.org/0000-0002-6663-8600"},"institutions":[{"id":"https://openalex.org/I74973139","display_name":"Carnegie Mellon University","ror":"https://ror.org/05x2bcf33","country_code":"US","type":"education","lineage":["https://openalex.org/I74973139"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Florian Metze","raw_affiliation_strings":["Carnegie Mellon University, Pittsburgh, PA, USA","Meta AI, USA","Technologies Institute, Carnegie Mellon University, USA"],"affiliations":[{"raw_affiliation_string":"Carnegie Mellon University, Pittsburgh, PA, USA","institution_ids":["https://openalex.org/I74973139"]},{"raw_affiliation_string":"Meta AI, USA","institution_ids":[]},{"raw_affiliation_string":"Technologies Institute, Carnegie Mellon University, USA","institution_ids":["https://openalex.org/I74973139"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5067919401","display_name":"Luke Zettlemoyer","orcid":"https://orcid.org/0009-0008-8296-0764"},"institutions":[{"id":"https://openalex.org/I4210114444","display_name":"Meta (United States)","ror":"https://ror.org/01zbnvs85","country_code":"US","type":"company","lineage":["https://openalex.org/I4210114444"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Luke Zettlemoyer","raw_affiliation_strings":["Meta Platforms Inc., Menlo Park, CA, USA"],"affiliations":[{"raw_affiliation_string":"Meta Platforms Inc., Menlo Park, CA, USA","institution_ids":["https://openalex.org/I4210114444"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5101170706","display_name":"Abdelrahman Mohamed","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Abdelrahman Mohamed","raw_affiliation_strings":["Rembrand Inc., Palo Alto, CA, USA","Meta AI, USA"],"affiliations":[{"raw_affiliation_string":"Rembrand Inc., Palo Alto, CA, USA","institution_ids":[]},{"raw_affiliation_string":"Meta AI, USA","institution_ids":[]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":8,"corresponding_author_ids":["https://openalex.org/A5068873086"],"corresponding_institution_ids":["https://openalex.org/I74973139"],"apc_list":null,"apc_paid":null,"fwci":2.0922,"has_fulltext":false,"cited_by_count":12,"citation_normalized_percentile":{"value":0.89528362,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":95,"max":98},"biblio":{"volume":"31","issue":null,"first_page":"3112","last_page":"3126"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10181","display_name":"Natural Language Processing Techniques","score":0.9998999834060669,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T10181","display_name":"Natural Language Processing Techniques","score":0.9998999834060669,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10201","display_name":"Speech Recognition and Synthesis","score":0.9998000264167786,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.9995999932289124,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.8155778646469116},{"id":"https://openalex.org/keywords/encoder","display_name":"Encoder","score":0.7470281720161438},{"id":"https://openalex.org/keywords/machine-translation","display_name":"Machine translation","score":0.521964967250824},{"id":"https://openalex.org/keywords/task","display_name":"Task (project management)","score":0.4998362064361572},{"id":"https://openalex.org/keywords/modular-design","display_name":"Modular design","score":0.4740214943885803},{"id":"https://openalex.org/keywords/software-portability","display_name":"Software portability","score":0.4696691930294037},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.4689077138900757},{"id":"https://openalex.org/keywords/interface","display_name":"Interface (matter)","score":0.41291549801826477},{"id":"https://openalex.org/keywords/speech-recognition","display_name":"Speech recognition","score":0.3980378210544586},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.33138149976730347},{"id":"https://openalex.org/keywords/programming-language","display_name":"Programming language","score":0.2075188159942627},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.15619084239006042}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8155778646469116},{"id":"https://openalex.org/C118505674","wikidata":"https://www.wikidata.org/wiki/Q42586063","display_name":"Encoder","level":2,"score":0.7470281720161438},{"id":"https://openalex.org/C203005215","wikidata":"https://www.wikidata.org/wiki/Q79798","display_name":"Machine translation","level":2,"score":0.521964967250824},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.4998362064361572},{"id":"https://openalex.org/C101468663","wikidata":"https://www.wikidata.org/wiki/Q1620158","display_name":"Modular design","level":2,"score":0.4740214943885803},{"id":"https://openalex.org/C63000827","wikidata":"https://www.wikidata.org/wiki/Q3080428","display_name":"Software portability","level":2,"score":0.4696691930294037},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.4689077138900757},{"id":"https://openalex.org/C113843644","wikidata":"https://www.wikidata.org/wiki/Q901882","display_name":"Interface (matter)","level":4,"score":0.41291549801826477},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.3980378210544586},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.33138149976730347},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.2075188159942627},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.15619084239006042},{"id":"https://openalex.org/C157915830","wikidata":"https://www.wikidata.org/wiki/Q2928001","display_name":"Bubble","level":2,"score":0.0},{"id":"https://openalex.org/C187736073","wikidata":"https://www.wikidata.org/wiki/Q2920921","display_name":"Management","level":1,"score":0.0},{"id":"https://openalex.org/C129307140","wikidata":"https://www.wikidata.org/wiki/Q6795880","display_name":"Maximum bubble pressure method","level":3,"score":0.0},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.0},{"id":"https://openalex.org/C162324750","wikidata":"https://www.wikidata.org/wiki/Q8134","display_name":"Economics","level":0,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/taslp.2023.3296019","is_oa":false,"landing_page_url":"https://doi.org/10.1109/taslp.2023.3296019","pdf_url":null,"source":{"id":"https://openalex.org/S4210169297","display_name":"IEEE/ACM Transactions on Audio Speech and Language Processing","issn_l":"2329-9290","issn":["2329-9290","2329-9304"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE/ACM Transactions on Audio, Speech, and Language Processing","raw_type":"journal-article"}],"best_oa_location":null,"sustainable_development_goals":[{"score":0.75,"display_name":"Quality Education","id":"https://metadata.un.org/sdg/4"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":93,"referenced_works":["https://openalex.org/W10259981","https://openalex.org/W1508165687","https://openalex.org/W1522301498","https://openalex.org/W1686946872","https://openalex.org/W1753482797","https://openalex.org/W1902568950","https://openalex.org/W1922655562","https://openalex.org/W2025653905","https://openalex.org/W2046932483","https://openalex.org/W2097333193","https://openalex.org/W2124033848","https://openalex.org/W2127141656","https://openalex.org/W2130942839","https://openalex.org/W2133564696","https://openalex.org/W2143612262","https://openalex.org/W2193413348","https://openalex.org/W2251321385","https://openalex.org/W2293634267","https://openalex.org/W2327501763","https://openalex.org/W2514741789","https://openalex.org/W2526425061","https://openalex.org/W2613904329","https://openalex.org/W2767206889","https://openalex.org/W2892213699","https://openalex.org/W2896457183","https://openalex.org/W2913988491","https://openalex.org/W2933138175","https://openalex.org/W2936774411","https://openalex.org/W2941814890","https://openalex.org/W2962780374","https://openalex.org/W2962784628","https://openalex.org/W2962826786","https://openalex.org/W2963211739","https://openalex.org/W2963216553","https://openalex.org/W2963250244","https://openalex.org/W2963394326","https://openalex.org/W2963434219","https://openalex.org/W2963532001","https://openalex.org/W2963807318","https://openalex.org/W2964118342","https://openalex.org/W2964262254","https://openalex.org/W2972818416","https://openalex.org/W2986385672","https://openalex.org/W2988975212","https://openalex.org/W2996251235","https://openalex.org/W3002595344","https://openalex.org/W3007328579","https://openalex.org/W3015162217","https://openalex.org/W3015698636","https://openalex.org/W3024605872","https://openalex.org/W3035445001","https://openalex.org/W3094502228","https://openalex.org/W3097747488","https://openalex.org/W3097777922","https://openalex.org/W3100753857","https://openalex.org/W3112157188","https://openalex.org/W3177172118","https://openalex.org/W4224308101","https://openalex.org/W4295312788","https://openalex.org/W4297669644","https://openalex.org/W4311000453","https://openalex.org/W4381786045","https://openalex.org/W4385245566","https://openalex.org/W4386566728","https://openalex.org/W6631190155","https://openalex.org/W6637108112","https://openalex.org/W6637698695","https://openalex.org/W6640090968","https://openalex.org/W6674571003","https://openalex.org/W6678360021","https://openalex.org/W6679434410","https://openalex.org/W6679436768","https://openalex.org/W6687566353","https://openalex.org/W6691770337","https://openalex.org/W6730038592","https://openalex.org/W6737778391","https://openalex.org/W6739901393","https://openalex.org/W6746208923","https://openalex.org/W6751744669","https://openalex.org/W6755207826","https://openalex.org/W6758771928","https://openalex.org/W6761563299","https://openalex.org/W6766978945","https://openalex.org/W6771862290","https://openalex.org/W6774835902","https://openalex.org/W6775291257","https://openalex.org/W6777028661","https://openalex.org/W6780226713","https://openalex.org/W6784333009","https://openalex.org/W6803034309","https://openalex.org/W6810081322","https://openalex.org/W6846002384","https://openalex.org/W6847363464"],"related_works":["https://openalex.org/W107105315","https://openalex.org/W1584537303","https://openalex.org/W4388155270","https://openalex.org/W1872724644","https://openalex.org/W2750549761","https://openalex.org/W28826848","https://openalex.org/W2122272819","https://openalex.org/W4367156293","https://openalex.org/W2130894091","https://openalex.org/W2994151208"],"abstract_inverted_index":{"State-of-the-art":[0],"encoder-decoder":[1,53],"models":[2,224,269],"(e.g.":[3],"for":[4,51,71,102,133,203],"machine":[5],"translation":[6],"(MT)":[7],"or":[8,215],"automatic":[9],"speech":[10],"recognition":[11],"(ASR))":[12],"are":[13],"constructed":[14],"and":[15,82,118,137,208,234],"trained":[16,186,268],"end-to-end":[17,267],"as":[18],"an":[19,276],"atomic":[20],"unit.":[21],"No":[22],"component":[23],"of":[24,90,112,127,151,171,182,219,280],"the":[25,31,69,78,110,115,119,125,166,180,204,209,217,226,241,248],"model":[26,256],"can":[27,62,197],"be":[28,63,198],"(re-)used":[29],"without":[30,68,200],"others,":[32],"making":[33],"it":[34],"impossible":[35],"to":[36,65,87,156,164,178],"share":[37],"parts,":[38],"e.g.":[39],"a":[40,49,56,88,94,145,152,185,253],"high":[41],"resourced":[42],"decoder,":[43],"across":[44,114,138],"tasks.":[45],"We":[46,98,174],"describe":[47],"LegoNN,":[48],"procedure":[50],"building":[52],"architectures":[54],"in":[55,162],"way":[57],"so":[58],"that":[59],"its":[60],"parts":[61],"applied":[64],"other":[66,120,139],"tasks":[67,132,140],"need":[70],"any":[72,201],"fine-tuning.":[73],"To":[74,123,245],"achieve":[75,235],"this":[76],"reusability,":[77],"interface":[79],"between":[80,130],"encoder":[81,148],"decoder":[83,128,190],"modules":[84,129,259],"is":[85,107,121],"grounded":[86],"sequence":[89],"marginal":[91],"distributions":[92],"over":[93],"pre-defined":[95],"discrete":[96],"vocabulary.":[97],"present":[99,175],"two":[100],"approaches":[101],"ingesting":[103],"these":[104],"marginals;":[105],"one":[106],"differentiable,":[108],"allowing":[109],"flow":[111],"gradients":[113],"entire":[116],"network,":[117],"gradient-isolating.":[122],"enable":[124],"portability":[126],"MT":[131,195,212,228],"different":[134,266,272],"source":[135],"languages":[136],"like":[141],"ASR,":[142],"we":[143,251],"introduce":[144],"modality":[146],"agnostic":[147],"which":[149],"consists":[150],"length":[153,169],"control":[154],"mechanism":[155],"dynamically":[157],"adapt":[158],"encoders'":[159],"output":[160],"lengths":[161],"order":[163],"match":[165],"expected":[167],"input":[168],"range":[170],"pre-trained":[172],"decoders.":[173],"several":[176],"experiments":[177],"demonstrate":[179],"effectiveness":[181],"LegoNN":[183,189,223,254],"models:":[184],"language":[187],"generation":[188],"module":[191],"from":[192,257],"German-English":[193],"(De-En)":[194],"task":[196,229],"reused":[199],"fine-tuning":[202],"Europarl":[205,242],"English":[206],"ASR":[207,243,255],"Romanian-English":[210],"(Ro-En)":[211],"tasks,":[213],"matching":[214],"beating":[216],"performance":[218],"baseline.":[220],"After":[221],"fine-tuning,":[222],"improve":[225],"Ro-En":[227],"by":[230],"1.5":[231],"BLEU":[232],"points":[233],"12.5%":[236],"relative":[237],"WER":[238,278],"reduction":[239,279],"on":[240,270],"task.":[244],"show":[246],"how":[247],"approach":[249],"generalizes,":[250],"compose":[252],"three":[258,271],"\u2013":[260,274],"each":[261],"has":[262],"been":[263],"learned":[264],"within":[265],"datasets":[273],"achieving":[275],"overall":[277],"19.5%.":[281]},"counts_by_year":[{"year":2025,"cited_by_count":2},{"year":2024,"cited_by_count":5},{"year":2023,"cited_by_count":5}],"updated_date":"2026-03-27T14:29:43.386196","created_date":"2025-10-10T00:00:00"}