{"id":"https://openalex.org/W4288086191","doi":"https://doi.org/10.1145/3357384.3358028","title":"How Does BERT Answer Questions?","display_name":"How Does BERT Answer Questions?","publication_year":2019,"publication_date":"2019-11-03","ids":{"openalex":"https://openalex.org/W4288086191","doi":"https://doi.org/10.1145/3357384.3358028"},"language":"en","primary_location":{"id":"doi:10.1145/3357384.3358028","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3357384.3358028","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 28th ACM International Conference on Information and Knowledge Management","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/A5077285800","display_name":"Betty van Aken","orcid":null},"institutions":[{"id":"https://openalex.org/I129643931","display_name":"Berliner Hochschule f\u00fcr Technik","ror":"https://ror.org/00w7whj55","country_code":"DE","type":"education","lineage":["https://openalex.org/I129643931"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Betty van Aken","raw_affiliation_strings":["Beuth University of Applied Sciences, Berlin, Germany"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Beuth University of Applied Sciences, Berlin, Germany","institution_ids":["https://openalex.org/I129643931"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5027602672","display_name":"Benjamin Winter","orcid":"https://orcid.org/0000-0003-4644-4099"},"institutions":[{"id":"https://openalex.org/I129643931","display_name":"Berliner Hochschule f\u00fcr Technik","ror":"https://ror.org/00w7whj55","country_code":"DE","type":"education","lineage":["https://openalex.org/I129643931"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Benjamin Winter","raw_affiliation_strings":["Beuth University of Applied Sciences, Berlin, Germany"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Beuth University of Applied Sciences, Berlin, Germany","institution_ids":["https://openalex.org/I129643931"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5082325279","display_name":"Alexander L\u00f6ser","orcid":"https://orcid.org/0000-0002-4440-3261"},"institutions":[{"id":"https://openalex.org/I129643931","display_name":"Berliner Hochschule f\u00fcr Technik","ror":"https://ror.org/00w7whj55","country_code":"DE","type":"education","lineage":["https://openalex.org/I129643931"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Alexander L\u00f6ser","raw_affiliation_strings":["Beuth University of Applied Sciences, Berlin, Germany"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Beuth University of Applied Sciences, Berlin, Germany","institution_ids":["https://openalex.org/I129643931"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5044994184","display_name":"Felix A. Gers","orcid":"https://orcid.org/0000-0002-2258-2983"},"institutions":[{"id":"https://openalex.org/I129643931","display_name":"Berliner Hochschule f\u00fcr Technik","ror":"https://ror.org/00w7whj55","country_code":"DE","type":"education","lineage":["https://openalex.org/I129643931"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Felix A. Gers","raw_affiliation_strings":["Beuth University of Applied Sciences, Berlin, Germany"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Beuth University of Applied Sciences, Berlin, Germany","institution_ids":["https://openalex.org/I129643931"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":4,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":4.1936,"has_fulltext":false,"cited_by_count":79,"citation_normalized_percentile":{"value":0.95376546,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":98,"max":100},"biblio":{"volume":null,"issue":null,"first_page":"1823","last_page":"1832"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":1.0,"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":1.0,"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":0.9983000159263611,"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.9965999722480774,"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":0.827686607837677},{"id":"https://openalex.org/keywords/security-token","display_name":"Security token","score":0.7715242505073547},{"id":"https://openalex.org/keywords/transformer","display_name":"Transformer","score":0.7130721211433411},{"id":"https://openalex.org/keywords/question-answering","display_name":"Question answering","score":0.6695910692214966},{"id":"https://openalex.org/keywords/encoder","display_name":"Encoder","score":0.6660138964653015},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.5452008247375488},{"id":"https://openalex.org/keywords/natural-language-processing","display_name":"Natural language processing","score":0.4775888919830322},{"id":"https://openalex.org/keywords/pipeline","display_name":"Pipeline (software)","score":0.4420109987258911},{"id":"https://openalex.org/keywords/representation","display_name":"Representation (politics)","score":0.4234040081501007},{"id":"https://openalex.org/keywords/task","display_name":"Task (project management)","score":0.42065584659576416},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.4178531765937805},{"id":"https://openalex.org/keywords/programming-language","display_name":"Programming language","score":0.12292984127998352}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.827686607837677},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.7715242505073547},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.7130721211433411},{"id":"https://openalex.org/C44291984","wikidata":"https://www.wikidata.org/wiki/Q1074173","display_name":"Question answering","level":2,"score":0.6695910692214966},{"id":"https://openalex.org/C118505674","wikidata":"https://www.wikidata.org/wiki/Q42586063","display_name":"Encoder","level":2,"score":0.6660138964653015},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5452008247375488},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.4775888919830322},{"id":"https://openalex.org/C43521106","wikidata":"https://www.wikidata.org/wiki/Q2165493","display_name":"Pipeline (software)","level":2,"score":0.4420109987258911},{"id":"https://openalex.org/C2776359362","wikidata":"https://www.wikidata.org/wiki/Q2145286","display_name":"Representation (politics)","level":3,"score":0.4234040081501007},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.42065584659576416},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.4178531765937805},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.12292984127998352},{"id":"https://openalex.org/C187736073","wikidata":"https://www.wikidata.org/wiki/Q2920921","display_name":"Management","level":1,"score":0.0},{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","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},{"id":"https://openalex.org/C62520636","wikidata":"https://www.wikidata.org/wiki/Q944","display_name":"Quantum mechanics","level":1,"score":0.0},{"id":"https://openalex.org/C94625758","wikidata":"https://www.wikidata.org/wiki/Q7163","display_name":"Politics","level":2,"score":0.0},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.0},{"id":"https://openalex.org/C165801399","wikidata":"https://www.wikidata.org/wiki/Q25428","display_name":"Voltage","level":2,"score":0.0},{"id":"https://openalex.org/C199539241","wikidata":"https://www.wikidata.org/wiki/Q7748","display_name":"Law","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},{"id":"https://openalex.org/C17744445","wikidata":"https://www.wikidata.org/wiki/Q36442","display_name":"Political science","level":0,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3357384.3358028","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3357384.3358028","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 28th ACM International Conference on Information and Knowledge Management","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[{"score":0.7799999713897705,"display_name":"Quality Education","id":"https://metadata.un.org/sdg/4"}],"awards":[{"id":"https://openalex.org/G755434222","display_name":null,"funder_award_id":"732328","funder_id":"https://openalex.org/F4320335254","funder_display_name":"Horizon 2020"}],"funders":[{"id":"https://openalex.org/F4320335254","display_name":"Horizon 2020","ror":null}],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":7,"referenced_works":["https://openalex.org/W2070246124","https://openalex.org/W2150593711","https://openalex.org/W2295676751","https://openalex.org/W2563574619","https://openalex.org/W2962739339","https://openalex.org/W2962772482","https://openalex.org/W2963323070"],"related_works":["https://openalex.org/W2384605597","https://openalex.org/W2387743295","https://openalex.org/W4388335561","https://openalex.org/W2970530566","https://openalex.org/W4288261899","https://openalex.org/W4307309205","https://openalex.org/W4288267738","https://openalex.org/W2964413124","https://openalex.org/W4388937922","https://openalex.org/W3113264705"],"abstract_inverted_index":{"Bidirectional":[0],"Encoder":[1],"Representations":[2],"from":[3],"Transformers":[4],"(BERT)":[5],"reach":[6],"state-of-the-art":[7],"results":[8,144],"in":[9,100,125,194],"a":[10,40,88,112],"variety":[11],"of":[12,19,43,80,87,114,132,198],"Natural":[13],"Language":[14],"Processing":[15],"tasks.":[16,160],"However,":[17],"understanding":[18],"their":[20,58],"internal":[21],"functioning":[22],"is":[23],"still":[24],"insufficient":[25],"and":[26,34,116,187],"unsatisfactory.":[27],"In":[28],"order":[29,101],"to":[30,102,157],"better":[31],"understand":[32],"BERT":[33,150],"other":[35],"Transformer-based":[36],"models,":[37],"we":[38,61,110],"present":[39],"layer-wise":[41],"analysis":[42,72,131,175],"BERT's":[44,140],"hidden":[45,64,133],"states.":[46],"Unlike":[47],"previous":[48],"research,":[49],"which":[50],"mainly":[51],"focuses":[52,73],"on":[53,74,77,182],"explaining":[54],"Transformer":[55],"models":[56,75,96],"by":[57],"attention":[59],"weights,":[60],"argue":[62],"that":[63,120,146,154,177,188],"states":[65],"contain":[66],"equally":[67],"valuable":[68],"information.":[69],"Specifically,":[70],"our":[71,174],"fine-tuned":[76],"the":[78,104,122,147,183,195],"task":[79],"Question":[81],"Answering":[82],"(QA)":[83],"as":[84],"an":[85],"example":[86],"complex":[89],"downstream":[90],"task.":[91],"We":[92],"inspect":[93],"how":[94],"QA":[95],"transform":[97],"token":[98,171],"vectors":[99],"find":[103],"correct":[105],"answer.":[106],"To":[107],"this":[108],"end,":[109],"apply":[111],"set":[113],"general":[115],"QA-specific":[117],"probing":[118],"tasks":[119],"reveal":[121],"information":[123,168],"stored":[124],"each":[126],"representation":[127],"layer.":[128],"Our":[129,143],"qualitative":[130],"state":[134],"visualizations":[135],"provides":[136],"additional":[137],"insights":[138],"into":[139,169],"reasoning":[141],"process.":[142],"show":[145],"transformations":[148],"within":[149],"go":[151],"through":[152],"phases":[153],"are":[155],"related":[156],"traditional":[158],"pipeline":[159],"The":[161],"system":[162],"can":[163,191],"therefore":[164],"implicitly":[165],"incorporate":[166],"task-specific":[167],"its":[170],"representations.":[172],"Furthermore,":[173],"reveals":[176],"fine-tuning":[178],"has":[179],"little":[180],"impact":[181],"models'":[184],"semantic":[185],"abilities":[186],"prediction":[189],"errors":[190],"be":[192],"recognized":[193],"vector":[196],"representations":[197],"even":[199],"early":[200],"layers.":[201]},"counts_by_year":[{"year":2026,"cited_by_count":3},{"year":2025,"cited_by_count":8},{"year":2024,"cited_by_count":18},{"year":2023,"cited_by_count":21},{"year":2022,"cited_by_count":11},{"year":2021,"cited_by_count":12},{"year":2020,"cited_by_count":6}],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2022-07-28T00:00:00"}