{"id":"https://openalex.org/W3097737911","doi":"https://doi.org/10.21437/interspeech.2020-1849","title":"Investigation of Large-Margin Softmax in Neural Language Modeling","display_name":"Investigation of Large-Margin Softmax in Neural Language Modeling","publication_year":2020,"publication_date":"2020-10-25","ids":{"openalex":"https://openalex.org/W3097737911","doi":"https://doi.org/10.21437/interspeech.2020-1849","mag":"3097737911"},"language":"en","primary_location":{"id":"doi:10.21437/interspeech.2020-1849","is_oa":false,"landing_page_url":"https://doi.org/10.21437/interspeech.2020-1849","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Interspeech 2020","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/A5056608595","display_name":"Jingjing Huo","orcid":"https://orcid.org/0000-0003-1733-8748"},"institutions":[{"id":"https://openalex.org/I887968799","display_name":"RWTH Aachen University","ror":"https://ror.org/04xfq0f34","country_code":"DE","type":"education","lineage":["https://openalex.org/I887968799"]}],"countries":["DE"],"is_corresponding":true,"raw_author_name":"Jingjing Huo","raw_affiliation_strings":["Human Language Technology and Pattern Recognition Group, Computer Science Department RWTH Aachen University, 52074 Aachen, Germany,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Human Language Technology and Pattern Recognition Group, Computer Science Department RWTH Aachen University, 52074 Aachen, Germany,","institution_ids":["https://openalex.org/I887968799"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5046577673","display_name":"Yingbo Gao","orcid":null},"institutions":[{"id":"https://openalex.org/I887968799","display_name":"RWTH Aachen University","ror":"https://ror.org/04xfq0f34","country_code":"DE","type":"education","lineage":["https://openalex.org/I887968799"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Yingbo Gao","raw_affiliation_strings":["AppTek GmbH, 52062 Aachen, Germany","Human Language Technology and Pattern Recognition Group, Computer Science Department RWTH Aachen University, 52074 Aachen, Germany,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"AppTek GmbH, 52062 Aachen, Germany","institution_ids":[]},{"raw_affiliation_string":"Human Language Technology and Pattern Recognition Group, Computer Science Department RWTH Aachen University, 52074 Aachen, Germany,","institution_ids":["https://openalex.org/I887968799"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5006737674","display_name":"Weiyue Wang","orcid":"https://orcid.org/0000-0002-9127-7544"},"institutions":[{"id":"https://openalex.org/I887968799","display_name":"RWTH Aachen University","ror":"https://ror.org/04xfq0f34","country_code":"DE","type":"education","lineage":["https://openalex.org/I887968799"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Weiyue Wang","raw_affiliation_strings":["Human Language Technology and Pattern Recognition Group, Computer Science Department RWTH Aachen University, 52074 Aachen, Germany,","AppTek GmbH, 52062 Aachen, Germany"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Human Language Technology and Pattern Recognition Group, Computer Science Department RWTH Aachen University, 52074 Aachen, Germany,","institution_ids":["https://openalex.org/I887968799"]},{"raw_affiliation_string":"AppTek GmbH, 52062 Aachen, Germany","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5088968292","display_name":"Ralf Schl\u00fcter","orcid":"https://orcid.org/0000-0003-2839-9247"},"institutions":[{"id":"https://openalex.org/I887968799","display_name":"RWTH Aachen University","ror":"https://ror.org/04xfq0f34","country_code":"DE","type":"education","lineage":["https://openalex.org/I887968799"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Ralf Schl\u00fcter","raw_affiliation_strings":["Human Language Technology and Pattern Recognition Group, Computer Science Department RWTH Aachen University, 52074 Aachen, Germany,","AppTek GmbH, 52062 Aachen, Germany"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Human Language Technology and Pattern Recognition Group, Computer Science Department RWTH Aachen University, 52074 Aachen, Germany,","institution_ids":["https://openalex.org/I887968799"]},{"raw_affiliation_string":"AppTek GmbH, 52062 Aachen, Germany","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5112501010","display_name":"Hermann Ney","orcid":null},"institutions":[{"id":"https://openalex.org/I887968799","display_name":"RWTH Aachen University","ror":"https://ror.org/04xfq0f34","country_code":"DE","type":"education","lineage":["https://openalex.org/I887968799"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Hermann Ney","raw_affiliation_strings":["Human Language Technology and Pattern Recognition Group, Computer Science Department RWTH Aachen University, 52074 Aachen, Germany,","AppTek GmbH, 52062 Aachen, Germany"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Human Language Technology and Pattern Recognition Group, Computer Science Department RWTH Aachen University, 52074 Aachen, Germany,","institution_ids":["https://openalex.org/I887968799"]},{"raw_affiliation_string":"AppTek GmbH, 52062 Aachen, Germany","institution_ids":[]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":5,"corresponding_author_ids":["https://openalex.org/A5056608595"],"corresponding_institution_ids":["https://openalex.org/I887968799"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.14428113,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"3645","last_page":"3649"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10201","display_name":"Speech Recognition and Synthesis","score":0.9991999864578247,"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":0.9991999864578247,"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":0.994700014591217,"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/T10601","display_name":"Handwritten Text Recognition Techniques","score":0.9944999814033508,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/margin","display_name":"Margin (machine learning)","score":0.7422176003456116},{"id":"https://openalex.org/keywords/softmax-function","display_name":"Softmax function","score":0.7260101437568665},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.6908204555511475},{"id":"https://openalex.org/keywords/natural-language-processing","display_name":"Natural language processing","score":0.5191568732261658},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.48802176117897034},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.45902520418167114},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.42940646409988403},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.19499757885932922}],"concepts":[{"id":"https://openalex.org/C774472","wikidata":"https://www.wikidata.org/wiki/Q6760393","display_name":"Margin (machine learning)","level":2,"score":0.7422176003456116},{"id":"https://openalex.org/C188441871","wikidata":"https://www.wikidata.org/wiki/Q7554146","display_name":"Softmax function","level":3,"score":0.7260101437568665},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6908204555511475},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.5191568732261658},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.48802176117897034},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.45902520418167114},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.42940646409988403},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.19499757885932922}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.21437/interspeech.2020-1849","is_oa":false,"landing_page_url":"https://doi.org/10.21437/interspeech.2020-1849","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Interspeech 2020","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[{"score":0.7400000095367432,"id":"https://metadata.un.org/sdg/10","display_name":"Reduced inequalities"}],"awards":[{"id":"https://openalex.org/G7366229540","display_name":"A Sequence Classification Framework for Human Language Technology","funder_award_id":"694537","funder_id":"https://openalex.org/F4320320300","funder_display_name":"European Commission"}],"funders":[{"id":"https://openalex.org/F4320320300","display_name":"European Commission","ror":"https://ror.org/00k4n6c32"}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":33,"referenced_works":["https://openalex.org/W30564313","https://openalex.org/W179875071","https://openalex.org/W1554981240","https://openalex.org/W1587353960","https://openalex.org/W1968625547","https://openalex.org/W1970689298","https://openalex.org/W1975846642","https://openalex.org/W1985258458","https://openalex.org/W2026149468","https://openalex.org/W2080213370","https://openalex.org/W2100664567","https://openalex.org/W2104997912","https://openalex.org/W2117553576","https://openalex.org/W2140679639","https://openalex.org/W2150142469","https://openalex.org/W2158148237","https://openalex.org/W2187089797","https://openalex.org/W2345190899","https://openalex.org/W2402268235","https://openalex.org/W2407031592","https://openalex.org/W2516310076","https://openalex.org/W2766557427","https://openalex.org/W2787642437","https://openalex.org/W2903357770","https://openalex.org/W2949007385","https://openalex.org/W2962898354","https://openalex.org/W2963466847","https://openalex.org/W2963656735","https://openalex.org/W2969985801","https://openalex.org/W2972692349","https://openalex.org/W3103152812","https://openalex.org/W4239510810","https://openalex.org/W4294149583"],"related_works":["https://openalex.org/W3107204728","https://openalex.org/W2982889384","https://openalex.org/W4226227567","https://openalex.org/W2971218105","https://openalex.org/W4287113729","https://openalex.org/W3173314472","https://openalex.org/W3103152812","https://openalex.org/W3170224572","https://openalex.org/W1773599773","https://openalex.org/W2268150819"],"abstract_inverted_index":{"To":[0],"encourage":[1],"intra-class":[2],"compactness":[3],"and":[4,16,45,59,80,92,134,184],"inter-class":[5],"separability":[6],"among":[7],"trainable":[8],"feature":[9],"vectors,":[10,179],"large-margin":[11,27,156],"softmax":[12,31,58,157,169],"methods":[13],"are":[14,65,173,187],"developed":[15],"widely":[17],"applied":[18],"in":[19,85,103,129,141],"the":[20,26,30,78,100,107,125,167,182],"face":[21,104],"recognition":[22],"community.The":[23],"introduction":[24],"of":[25,96,109,166,177],"concept":[28],"into":[29],"is":[32,51,149],"reported":[33],"to":[34,67,72,164],"have":[35],"good":[36],"properties":[37],"such":[38],"as":[39],"enhanced":[40],"discriminative":[41],"power,":[42],"less":[43],"overfitting":[44],"welldefined":[46],"geometric":[47],"intuitions.Nowadays,":[48],"language":[49,74,111,137,153],"modeling":[50],"commonly":[52],"approached":[53],"with":[54,131,155],"neural":[55,73,136,152],"networks":[56],"using":[57],"cross":[60],"entropy.In":[61],"this":[62],"work,":[63],"we":[64,89,113,123],"curious":[66],"see":[68],"if":[69],"introducing":[70],"large-margins":[71],"models":[75,138,154],"would":[76],"improve":[77],"perplexity":[79,148],"consequently":[81],"word":[82,119,160,178],"error":[83,161],"rate":[84,162],"automatic":[86,142],"speech":[87,143],"recognition.Specifically,":[88],"first":[90],"implement":[91],"test":[93],"various":[94,132],"types":[95],"conventional":[97],"margins":[98,133,172],"following":[99],"previous":[101],"works":[102],"recognition.To":[105],"address":[106],"distribution":[108],"natural":[110],"data,":[112],"then":[114],"compare":[115],"different":[116],"strategies":[117],"for":[118],"vector":[120],"normscaling.After":[121],"that,":[122],"apply":[124],"best":[126],"norm-scaling":[127],"setup":[128],"combination":[130],"conduct":[135],"rescoring":[139],"experiments":[140],"recognition.We":[144],"find":[145],"that":[146,165,181],"although":[147],"slightly":[150],"deteriorated,":[151],"can":[158],"yield":[159],"similar":[163],"standard":[168],"baseline.Finally,":[170],"expected":[171],"analyzed":[174],"through":[175],"visualization":[176],"showing":[180],"syntactic":[183],"semantic":[185],"relationships":[186],"also":[188],"preserved.":[189]},"counts_by_year":[],"updated_date":"2026-05-23T08:51:43.019350","created_date":"2025-10-10T00:00:00"}