{"id":"https://openalex.org/W2932970397","doi":"https://doi.org/10.1109/icassp.2019.8683533","title":"Knowledge Distillation for Recurrent Neural Network Language Modeling with Trust Regularization","display_name":"Knowledge Distillation for Recurrent Neural Network Language Modeling with Trust Regularization","publication_year":2019,"publication_date":"2019-04-17","ids":{"openalex":"https://openalex.org/W2932970397","doi":"https://doi.org/10.1109/icassp.2019.8683533","mag":"2932970397"},"language":"en","primary_location":{"id":"doi:10.1109/icassp.2019.8683533","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp.2019.8683533","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2019 - 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","raw_type":"proceedings-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/A5103247973","display_name":"Yangyang Shi","orcid":"https://orcid.org/0000-0001-5297-4155"},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Yangyang Shi","raw_affiliation_strings":["Mobvoi AI Lab"],"affiliations":[{"raw_affiliation_string":"Mobvoi AI Lab","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5112467501","display_name":"Mei-Yuh Hwang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Mei-Yuh Hwang","raw_affiliation_strings":["Mobvoi AI Lab"],"affiliations":[{"raw_affiliation_string":"Mobvoi AI Lab","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5049457213","display_name":"Xin Lei","orcid":"https://orcid.org/0000-0002-4400-4068"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xin Lei","raw_affiliation_strings":["Mobvoi AI Lab"],"affiliations":[{"raw_affiliation_string":"Mobvoi AI Lab","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5008349950","display_name":"Haoyu Sheng","orcid":null},"institutions":[{"id":"https://openalex.org/I22675022","display_name":"Williams College","ror":"https://ror.org/04avkmd49","country_code":"US","type":"education","lineage":["https://openalex.org/I22675022"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Haoyu Sheng","raw_affiliation_strings":["Williams College"],"affiliations":[{"raw_affiliation_string":"Williams College","institution_ids":["https://openalex.org/I22675022"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":4,"corresponding_author_ids":["https://openalex.org/A5103247973"],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":3.0803,"has_fulltext":false,"cited_by_count":24,"citation_normalized_percentile":{"value":0.9329241,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":89,"max":99},"biblio":{"volume":null,"issue":null,"first_page":"7230","last_page":"7234"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","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/T10028","display_name":"Topic Modeling","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/T10181","display_name":"Natural Language Processing Techniques","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/T10201","display_name":"Speech Recognition and Synthesis","score":0.9976999759674072,"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/perplexity","display_name":"Perplexity","score":0.879912257194519},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.8602286577224731},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.8580513000488281},{"id":"https://openalex.org/keywords/treebank","display_name":"Treebank","score":0.7751801013946533},{"id":"https://openalex.org/keywords/regularization","display_name":"Regularization (linguistics)","score":0.6503239274024963},{"id":"https://openalex.org/keywords/distillation","display_name":"Distillation","score":0.6277765035629272},{"id":"https://openalex.org/keywords/recurrent-neural-network","display_name":"Recurrent neural network","score":0.6243997812271118},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.5699496269226074},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.5286398530006409},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.5179230570793152},{"id":"https://openalex.org/keywords/word-error-rate","display_name":"Word error rate","score":0.5034882426261902},{"id":"https://openalex.org/keywords/test-set","display_name":"Test set","score":0.42405885457992554},{"id":"https://openalex.org/keywords/natural-language-processing","display_name":"Natural language processing","score":0.33315569162368774}],"concepts":[{"id":"https://openalex.org/C100279451","wikidata":"https://www.wikidata.org/wiki/Q372193","display_name":"Perplexity","level":3,"score":0.879912257194519},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.8602286577224731},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8580513000488281},{"id":"https://openalex.org/C206134035","wikidata":"https://www.wikidata.org/wiki/Q811525","display_name":"Treebank","level":3,"score":0.7751801013946533},{"id":"https://openalex.org/C2776135515","wikidata":"https://www.wikidata.org/wiki/Q17143721","display_name":"Regularization (linguistics)","level":2,"score":0.6503239274024963},{"id":"https://openalex.org/C204030448","wikidata":"https://www.wikidata.org/wiki/Q101017","display_name":"Distillation","level":2,"score":0.6277765035629272},{"id":"https://openalex.org/C147168706","wikidata":"https://www.wikidata.org/wiki/Q1457734","display_name":"Recurrent neural network","level":3,"score":0.6243997812271118},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5699496269226074},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5286398530006409},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.5179230570793152},{"id":"https://openalex.org/C40969351","wikidata":"https://www.wikidata.org/wiki/Q3516228","display_name":"Word error rate","level":2,"score":0.5034882426261902},{"id":"https://openalex.org/C169903167","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Test set","level":2,"score":0.42405885457992554},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.33315569162368774},{"id":"https://openalex.org/C178790620","wikidata":"https://www.wikidata.org/wiki/Q11351","display_name":"Organic chemistry","level":1,"score":0.0},{"id":"https://openalex.org/C185592680","wikidata":"https://www.wikidata.org/wiki/Q2329","display_name":"Chemistry","level":0,"score":0.0},{"id":"https://openalex.org/C19768560","wikidata":"https://www.wikidata.org/wiki/Q320727","display_name":"Dependency (UML)","level":2,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/icassp.2019.8683533","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp.2019.8683533","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2019 - 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[{"id":"https://metadata.un.org/sdg/16","display_name":"Peace, Justice and strong institutions","score":0.7300000190734863}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":54,"referenced_works":["https://openalex.org/W4919037","https://openalex.org/W179875071","https://openalex.org/W1524333225","https://openalex.org/W1591801644","https://openalex.org/W1690739335","https://openalex.org/W1821462560","https://openalex.org/W1938755728","https://openalex.org/W1999965501","https://openalex.org/W2024490156","https://openalex.org/W2161591461","https://openalex.org/W2212703438","https://openalex.org/W2250379827","https://openalex.org/W2286365479","https://openalex.org/W2294543795","https://openalex.org/W2549416390","https://openalex.org/W2553303224","https://openalex.org/W2571859396","https://openalex.org/W2951559648","https://openalex.org/W2951672049","https://openalex.org/W2952899695","https://openalex.org/W2962832505","https://openalex.org/W2962964385","https://openalex.org/W2962965405","https://openalex.org/W2963266340","https://openalex.org/W2963374479","https://openalex.org/W2963494889","https://openalex.org/W2963537482","https://openalex.org/W2963643655","https://openalex.org/W2963748792","https://openalex.org/W2963983719","https://openalex.org/W2964118293","https://openalex.org/W4294555862","https://openalex.org/W4298422451","https://openalex.org/W4299838440","https://openalex.org/W4300427683","https://openalex.org/W6600213771","https://openalex.org/W6631362777","https://openalex.org/W6635446068","https://openalex.org/W6637551013","https://openalex.org/W6638523607","https://openalex.org/W6640598943","https://openalex.org/W6683826617","https://openalex.org/W6688167117","https://openalex.org/W6691752037","https://openalex.org/W6696004547","https://openalex.org/W6696912008","https://openalex.org/W6720905350","https://openalex.org/W6727099177","https://openalex.org/W6729239390","https://openalex.org/W6729956949","https://openalex.org/W6731780175","https://openalex.org/W6741459021","https://openalex.org/W6742632731","https://openalex.org/W6745265922"],"related_works":["https://openalex.org/W4301800915","https://openalex.org/W2739034105","https://openalex.org/W2963839582","https://openalex.org/W4299838440","https://openalex.org/W2964047924","https://openalex.org/W2743945814","https://openalex.org/W2962832505","https://openalex.org/W2050138804","https://openalex.org/W2151348424","https://openalex.org/W1494910745"],"abstract_inverted_index":{"Recurrent":[0,23],"Neural":[1,24],"Networks":[2],"(RNNs)":[3],"have":[4,39],"dominated":[5],"language":[6,26],"modeling":[7],"because":[8],"of":[9,32,54,95,97,128],"their":[10],"superior":[11],"performance":[12,140],"over":[13],"traditional":[14],"N-gram":[15],"based":[16],"models.":[17],"In":[18,47,65,113],"many":[19],"applications,":[20],"a":[21,69,93,114],"large":[22,40],"Network":[25],"model":[27,61,102,124],"(RNNLM)":[28],"or":[29],"an":[30],"ensemble":[31],"several":[33],"RNNLMs":[34],"is":[35],"used.":[36],"These":[37],"models":[38],"memory":[41],"footprints":[42],"and":[43],"require":[44],"heavy":[45],"computation.":[46],"this":[48],"paper,":[49],"we":[50,67,87,120],"examine":[51],"the":[52,60,75,89,98,105,122,129],"effect":[53],"applying":[55],"knowledge":[56,76,82],"distillation":[57,77,83],"in":[58,135],"reducing":[59],"size":[62,91,125],"for":[63,79],"RNNLMs.":[64,80],"addition,":[66],"propose":[68],"trust":[70,85],"regularization":[71],"method":[72],"to":[73,92,126],"improve":[74],"training":[78],"Using":[81],"with":[84,132],"regularization,":[86],"reduce":[88,121],"parameter":[90],"third":[94],"that":[96],"previously":[99],"published":[100],"best":[101],"while":[103],"maintaining":[104],"state-of-the-art":[106],"perplexity":[107],"result":[108],"on":[109,141],"Penn":[110],"Treebank":[111],"data.":[112],"speech":[115],"recognition":[116],"N-best":[117],"rescoring":[118],"task,":[119],"RNNLM":[123],"18.5%":[127],"baseline":[130],"system,":[131],"no":[133],"degradation":[134],"word":[136],"error":[137],"rate":[138],"(WER)":[139],"Wall":[142],"Street":[143],"Journal":[144],"data":[145],"set.":[146]},"counts_by_year":[{"year":2025,"cited_by_count":1},{"year":2023,"cited_by_count":1},{"year":2022,"cited_by_count":2},{"year":2021,"cited_by_count":8},{"year":2020,"cited_by_count":9},{"year":2019,"cited_by_count":3}],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-10T00:00:00"}