{"id":"https://openalex.org/W2996834012","doi":"https://doi.org/10.1145/3336191.3371792","title":"Model Compression with Two-stage Multi-teacher Knowledge Distillation for Web Question Answering System","display_name":"Model Compression with Two-stage Multi-teacher Knowledge Distillation for Web Question Answering System","publication_year":2020,"publication_date":"2020-01-20","ids":{"openalex":"https://openalex.org/W2996834012","doi":"https://doi.org/10.1145/3336191.3371792","mag":"2996834012"},"language":"en","primary_location":{"id":"doi:10.1145/3336191.3371792","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3336191.3371792","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 13th International Conference on Web Search and Data Mining","raw_type":"proceedings-article"},"type":"preprint","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/A5100758950","display_name":"Ze Yang","orcid":"https://orcid.org/0000-0003-3786-8842"},"institutions":[{"id":"https://openalex.org/I4210113369","display_name":"Microsoft Research Asia (China)","ror":"https://ror.org/0300m5276","country_code":"CN","type":"company","lineage":["https://openalex.org/I1290206253","https://openalex.org/I4210113369"]}],"countries":["CN"],"is_corresponding":true,"raw_author_name":"Ze Yang","raw_affiliation_strings":["Microsoft, Beijing, China"],"affiliations":[{"raw_affiliation_string":"Microsoft, Beijing, China","institution_ids":["https://openalex.org/I4210113369"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5077262995","display_name":"Linjun Shou","orcid":"https://orcid.org/0000-0002-1050-7708"},"institutions":[{"id":"https://openalex.org/I4210113369","display_name":"Microsoft Research Asia (China)","ror":"https://ror.org/0300m5276","country_code":"CN","type":"company","lineage":["https://openalex.org/I1290206253","https://openalex.org/I4210113369"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Linjun Shou","raw_affiliation_strings":["Microsoft, Beijing, China"],"affiliations":[{"raw_affiliation_string":"Microsoft, Beijing, China","institution_ids":["https://openalex.org/I4210113369"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101640051","display_name":"Ming Gong","orcid":"https://orcid.org/0000-0001-6140-7187"},"institutions":[{"id":"https://openalex.org/I4210113369","display_name":"Microsoft Research Asia (China)","ror":"https://ror.org/0300m5276","country_code":"CN","type":"company","lineage":["https://openalex.org/I1290206253","https://openalex.org/I4210113369"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Ming Gong","raw_affiliation_strings":["Microsoft, Beijing, China"],"affiliations":[{"raw_affiliation_string":"Microsoft, Beijing, China","institution_ids":["https://openalex.org/I4210113369"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5090739176","display_name":"Wutao Lin","orcid":null},"institutions":[{"id":"https://openalex.org/I4210113369","display_name":"Microsoft Research Asia (China)","ror":"https://ror.org/0300m5276","country_code":"CN","type":"company","lineage":["https://openalex.org/I1290206253","https://openalex.org/I4210113369"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Wutao Lin","raw_affiliation_strings":["Microsoft, Beijing, China"],"affiliations":[{"raw_affiliation_string":"Microsoft, Beijing, China","institution_ids":["https://openalex.org/I4210113369"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5060116992","display_name":"Daxin Jiang","orcid":"https://orcid.org/0000-0002-6657-5806"},"institutions":[{"id":"https://openalex.org/I4210113369","display_name":"Microsoft Research Asia (China)","ror":"https://ror.org/0300m5276","country_code":"CN","type":"company","lineage":["https://openalex.org/I1290206253","https://openalex.org/I4210113369"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Daxin Jiang","raw_affiliation_strings":["Microsoft, Beijing, China"],"affiliations":[{"raw_affiliation_string":"Microsoft, Beijing, China","institution_ids":["https://openalex.org/I4210113369"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":5,"corresponding_author_ids":["https://openalex.org/A5100758950"],"corresponding_institution_ids":["https://openalex.org/I4210113369"],"apc_list":null,"apc_paid":null,"fwci":9.3159,"has_fulltext":false,"cited_by_count":92,"citation_normalized_percentile":{"value":0.98347675,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":97,"max":100},"biblio":{"volume":null,"issue":null,"first_page":"690","last_page":"698"},"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":0.9957000017166138,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.9951000213623047,"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/overfitting","display_name":"Overfitting","score":0.8979475498199463},{"id":"https://openalex.org/keywords/question-answering","display_name":"Question answering","score":0.7672383189201355},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.755650520324707},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.754729688167572},{"id":"https://openalex.org/keywords/distillation","display_name":"Distillation","score":0.7063805460929871},{"id":"https://openalex.org/keywords/task","display_name":"Task (project management)","score":0.6401168704032898},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.5852754712104797},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.5474109053611755},{"id":"https://openalex.org/keywords/speedup","display_name":"Speedup","score":0.5389407873153687},{"id":"https://openalex.org/keywords/baseline","display_name":"Baseline (sea)","score":0.5019021034240723},{"id":"https://openalex.org/keywords/natural-language-processing","display_name":"Natural language processing","score":0.3430272340774536},{"id":"https://openalex.org/keywords/information-retrieval","display_name":"Information retrieval","score":0.32817715406417847},{"id":"https://openalex.org/keywords/engineering","display_name":"Engineering","score":0.09019911289215088}],"concepts":[{"id":"https://openalex.org/C22019652","wikidata":"https://www.wikidata.org/wiki/Q331309","display_name":"Overfitting","level":3,"score":0.8979475498199463},{"id":"https://openalex.org/C44291984","wikidata":"https://www.wikidata.org/wiki/Q1074173","display_name":"Question answering","level":2,"score":0.7672383189201355},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.755650520324707},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.754729688167572},{"id":"https://openalex.org/C204030448","wikidata":"https://www.wikidata.org/wiki/Q101017","display_name":"Distillation","level":2,"score":0.7063805460929871},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.6401168704032898},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5852754712104797},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5474109053611755},{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.5389407873153687},{"id":"https://openalex.org/C12725497","wikidata":"https://www.wikidata.org/wiki/Q810247","display_name":"Baseline (sea)","level":2,"score":0.5019021034240723},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.3430272340774536},{"id":"https://openalex.org/C23123220","wikidata":"https://www.wikidata.org/wiki/Q816826","display_name":"Information retrieval","level":1,"score":0.32817715406417847},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.09019911289215088},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.0},{"id":"https://openalex.org/C201995342","wikidata":"https://www.wikidata.org/wiki/Q682496","display_name":"Systems engineering","level":1,"score":0.0},{"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/C127313418","wikidata":"https://www.wikidata.org/wiki/Q1069","display_name":"Geology","level":0,"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/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.0},{"id":"https://openalex.org/C111368507","wikidata":"https://www.wikidata.org/wiki/Q43518","display_name":"Oceanography","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3336191.3371792","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3336191.3371792","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 13th International Conference on Web Search and Data Mining","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[{"id":"https://metadata.un.org/sdg/4","display_name":"Quality Education","score":0.6499999761581421}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":24,"referenced_works":["https://openalex.org/W591148856","https://openalex.org/W1821462560","https://openalex.org/W1902041153","https://openalex.org/W1996901117","https://openalex.org/W2065130322","https://openalex.org/W2117130368","https://openalex.org/W2125389748","https://openalex.org/W2172734211","https://openalex.org/W2594843958","https://openalex.org/W2743289088","https://openalex.org/W2787752464","https://openalex.org/W2805003733","https://openalex.org/W2890407764","https://openalex.org/W2896457183","https://openalex.org/W2930786691","https://openalex.org/W2937297214","https://openalex.org/W2950967261","https://openalex.org/W2962739339","https://openalex.org/W2963213121","https://openalex.org/W2963736842","https://openalex.org/W4206070857","https://openalex.org/W4211165624","https://openalex.org/W4236965008","https://openalex.org/W6677103964"],"related_works":["https://openalex.org/W2989932438","https://openalex.org/W2795435272","https://openalex.org/W2951851447","https://openalex.org/W3099765033","https://openalex.org/W1893928041","https://openalex.org/W2753840555","https://openalex.org/W2051167396","https://openalex.org/W4210794429","https://openalex.org/W2996834012","https://openalex.org/W2980357766"],"abstract_inverted_index":{"Deep":[0],"pre-training":[1],"and":[2,8,108,142,164],"fine-tuning":[3],"models":[4,33,42,70],"(such":[5],"as":[6],"BERT":[7],"OpenAI":[9],"GPT)":[10],"have":[11],"demonstrated":[12],"excellent":[13],"results":[14,153,168],"in":[15,138],"question":[16],"answering":[17],"areas.":[18],"However,":[19],"due":[20],"to":[21,38,43,68,147],"the":[22,28,63,73,135,148,161,170],"sheer":[23],"amount":[24],"of":[25,31,178],"model":[26,54,64,106,114,179],"parameters,":[27],"inference":[29],"speed":[30],"these":[32,40],"is":[34],"very":[35],"slow.":[36],"How":[37],"apply":[39],"complex":[41],"real":[44],"business":[45],"scenarios":[46],"becomes":[47],"a":[48,82,99],"challenging":[49],"but":[50],"practical":[51],"problem.":[52],"Previous":[53],"compression":[55,65],"methods":[56,163],"usually":[57],"suffer":[58],"from":[59,130],"information":[60],"loss":[61],"during":[62],"procedure,":[66],"leading":[67],"inferior":[69],"compared":[71],"with":[72,115,169,175],"original":[74,171],"one.":[75],"To":[76],"tackle":[77],"this":[78,111],"challenge,":[79],"we":[80],"propose":[81],"Two-stage":[83],"Multi-teacher":[84],"Knowledge":[85],"Distillation":[86],"(TMKD":[87],"for":[88,91,104],"short)":[89],"method":[90,157],"web":[92],"Question":[93],"Answering":[94],"system.":[95],"We":[96],"first":[97],"develop":[98],"general":[100,145],"Q&A":[101,124],"distillation":[102,118],"task":[103],"student":[105,113,149],"pre-training,":[107],"further":[109],"fine-tune":[110],"pre-trained":[112],"multi-teacher":[116],"knowledge":[117,146],"on":[119],"downstream":[120],"tasks":[121,129],"(like":[122],"Web":[123],"task,":[125],"MNLI,":[126],"SNLI,":[127],"RTE":[128],"GLUE),":[131],"which":[132],"effectively":[133],"reduces":[134],"overfitting":[136],"bias":[137],"individual":[139],"teacher":[140,172],"models,":[141,173],"transfers":[143],"more":[144],"model.":[150],"The":[151],"experiment":[152],"show":[154],"that":[155],"our":[156],"can":[158],"significantly":[159],"outperform":[160],"baseline":[162],"even":[165],"achieve":[166],"comparable":[167],"along":[174],"substantial":[176],"speedup":[177],"inference.":[180]},"counts_by_year":[{"year":2026,"cited_by_count":1},{"year":2025,"cited_by_count":9},{"year":2024,"cited_by_count":14},{"year":2023,"cited_by_count":18},{"year":2022,"cited_by_count":18},{"year":2021,"cited_by_count":23},{"year":2020,"cited_by_count":9}],"updated_date":"2026-03-08T08:50:53.379069","created_date":"2025-10-10T00:00:00"}