{"id":"https://openalex.org/W3161911362","doi":"https://doi.org/10.1109/icassp39728.2021.9414724","title":"Large Margin Training Improves Language Models for ASR","display_name":"Large Margin Training Improves Language Models for ASR","publication_year":2021,"publication_date":"2021-05-13","ids":{"openalex":"https://openalex.org/W3161911362","doi":"https://doi.org/10.1109/icassp39728.2021.9414724","mag":"3161911362"},"language":"en","primary_location":{"id":"doi:10.1109/icassp39728.2021.9414724","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp39728.2021.9414724","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2021 - 2021 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/A5100781745","display_name":"Jilin Wang","orcid":"https://orcid.org/0009-0007-5762-3144"},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Jilin Wang","raw_affiliation_strings":["Baidu Research, Sunnyvale, CA, USA","Boston University, Boston, MA, USA"],"affiliations":[{"raw_affiliation_string":"Baidu Research, Sunnyvale, CA, USA","institution_ids":[]},{"raw_affiliation_string":"Boston University, Boston, MA, USA","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5082470433","display_name":"Jiaji Huang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jiaji Huang","raw_affiliation_strings":["Baidu Research, Sunnyvale, CA, USA"],"affiliations":[{"raw_affiliation_string":"Baidu Research, Sunnyvale, CA, USA","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5016543371","display_name":"Kenneth Church","orcid":"https://orcid.org/0000-0001-8378-6069"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Kenneth Ward Church","raw_affiliation_strings":["Baidu Research, Sunnyvale, CA, USA"],"affiliations":[{"raw_affiliation_string":"Baidu Research, Sunnyvale, CA, USA","institution_ids":[]}]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":3,"corresponding_author_ids":["https://openalex.org/A5100781745"],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.2719,"has_fulltext":false,"cited_by_count":2,"citation_normalized_percentile":{"value":0.61410863,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":93,"max":95},"biblio":{"volume":"1","issue":null,"first_page":"7368","last_page":"7372"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10201","display_name":"Speech Recognition and Synthesis","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/T10201","display_name":"Speech Recognition and Synthesis","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":0.9993000030517578,"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/T12031","display_name":"Speech and dialogue systems","score":0.9977999925613403,"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":0.9839973449707031},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.797305703163147},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.7732155323028564},{"id":"https://openalex.org/keywords/encoder","display_name":"Encoder","score":0.7526959180831909},{"id":"https://openalex.org/keywords/margin","display_name":"Margin (machine learning)","score":0.7515842914581299},{"id":"https://openalex.org/keywords/transformer","display_name":"Transformer","score":0.7410709261894226},{"id":"https://openalex.org/keywords/speech-recognition","display_name":"Speech recognition","score":0.5407860279083252},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.5164037346839905},{"id":"https://openalex.org/keywords/natural-language-processing","display_name":"Natural language processing","score":0.3762325048446655},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.30900537967681885},{"id":"https://openalex.org/keywords/voltage","display_name":"Voltage","score":0.10861939191818237},{"id":"https://openalex.org/keywords/engineering","display_name":"Engineering","score":0.07884740829467773}],"concepts":[{"id":"https://openalex.org/C100279451","wikidata":"https://www.wikidata.org/wiki/Q372193","display_name":"Perplexity","level":3,"score":0.9839973449707031},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.797305703163147},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.7732155323028564},{"id":"https://openalex.org/C118505674","wikidata":"https://www.wikidata.org/wiki/Q42586063","display_name":"Encoder","level":2,"score":0.7526959180831909},{"id":"https://openalex.org/C774472","wikidata":"https://www.wikidata.org/wiki/Q6760393","display_name":"Margin (machine learning)","level":2,"score":0.7515842914581299},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.7410709261894226},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.5407860279083252},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5164037346839905},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.3762325048446655},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.30900537967681885},{"id":"https://openalex.org/C165801399","wikidata":"https://www.wikidata.org/wiki/Q25428","display_name":"Voltage","level":2,"score":0.10861939191818237},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.07884740829467773},{"id":"https://openalex.org/C119599485","wikidata":"https://www.wikidata.org/wiki/Q43035","display_name":"Electrical engineering","level":1,"score":0.0},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/icassp39728.2021.9414724","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp39728.2021.9414724","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[{"display_name":"Reduced inequalities","score":0.6399999856948853,"id":"https://metadata.un.org/sdg/10"}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":37,"referenced_works":["https://openalex.org/W1494198834","https://openalex.org/W1524333225","https://openalex.org/W1566289585","https://openalex.org/W1566877644","https://openalex.org/W2020073413","https://openalex.org/W2064675550","https://openalex.org/W2080213370","https://openalex.org/W2120498807","https://openalex.org/W2158148237","https://openalex.org/W2158195707","https://openalex.org/W2747917286","https://openalex.org/W2787642437","https://openalex.org/W2888867175","https://openalex.org/W2896457183","https://openalex.org/W2923014074","https://openalex.org/W2963048539","https://openalex.org/W2963310665","https://openalex.org/W2963341956","https://openalex.org/W2963403868","https://openalex.org/W2964013229","https://openalex.org/W2972818416","https://openalex.org/W2990391581","https://openalex.org/W2997200074","https://openalex.org/W3007328579","https://openalex.org/W3027083471","https://openalex.org/W3034775979","https://openalex.org/W3088059392","https://openalex.org/W3096338464","https://openalex.org/W3103005696","https://openalex.org/W4294149583","https://openalex.org/W4385245566","https://openalex.org/W6631362777","https://openalex.org/W6739901393","https://openalex.org/W6749464311","https://openalex.org/W6755207826","https://openalex.org/W6770528390","https://openalex.org/W6779431784"],"related_works":["https://openalex.org/W2169518243","https://openalex.org/W2252095989","https://openalex.org/W2551914602","https://openalex.org/W4322096525","https://openalex.org/W2020757772","https://openalex.org/W4281893144","https://openalex.org/W2105076537","https://openalex.org/W3013624417","https://openalex.org/W4287826556","https://openalex.org/W3049463507"],"abstract_inverted_index":{"Language":[0],"models":[1],"(LM)":[2],"have":[3],"been":[4],"widely":[5],"deployed":[6],"in":[7],"modern":[8],"ASR":[9],"systems.":[10],"The":[11,72],"LM":[12,54,80],"is":[13,47,75],"often":[14],"trained":[15,100],"by":[16,101],"minimizing":[17,102],"its":[18],"perplexity":[19],"on":[20],"speech":[21],"transcript.":[22],"However,":[23],"few":[24],"studies":[25],"try":[26],"to":[27,55,60,68,77],"discriminate":[28],"a":[29,40,48,57,65],"\"gold\"":[30,62],"reference":[31],"against":[32],"inferior":[33,70],"hypotheses.":[34],"In":[35],"this":[36],"work,":[37],"we":[38],"propose":[39],"large":[41],"margin":[42],"language":[43],"model":[44],"(LMLM).":[45],"LMLM":[46,92],"general":[49,73],"framework":[50,74],"that":[51,91,98],"enforces":[52],"an":[53],"assign":[56],"higher":[58],"score":[59],"the":[61,69,111,117],"reference,":[63],"and":[64,86],"lower":[66],"one":[67],"hypothesis.":[71],"applied":[76],"three":[78,112],"pretrained":[79],"architectures:":[81],"left-to-right":[82],"LSTM,":[83],"transformer":[84,87,114],"encoder,":[85],"decoder.":[88],"Results":[89],"show":[90],"can":[93],"significantly":[94],"outperform":[95],"traditional":[96],"LMs":[97],"are":[99],"perplexity.":[103],"Especially":[104],"for":[105],"challenging":[106],"noisy":[107],"cases.":[108],"Finally,":[109],"among":[110],"architectures,":[113],"encoder":[115],"achieves":[116],"best":[118],"performance.":[119]},"counts_by_year":[{"year":2021,"cited_by_count":2}],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-10T00:00:00"}