{"id":"https://openalex.org/W4385768057","doi":"https://doi.org/10.24963/ijcai.2023/563","title":"SmartBERT: A Promotion of Dynamic Early Exiting Mechanism for Accelerating BERT Inference","display_name":"SmartBERT: A Promotion of Dynamic Early Exiting Mechanism for Accelerating BERT Inference","publication_year":2023,"publication_date":"2023-08-01","ids":{"openalex":"https://openalex.org/W4385768057","doi":"https://doi.org/10.24963/ijcai.2023/563"},"language":"en","primary_location":{"id":"doi:10.24963/ijcai.2023/563","is_oa":true,"landing_page_url":"https://doi.org/10.24963/ijcai.2023/563","pdf_url":"https://www.ijcai.org/proceedings/2023/0563.pdf","source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the Thirty-Second International Joint Conference on Artificial Intelligence","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://www.ijcai.org/proceedings/2023/0563.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5058039212","display_name":"Boren Hu","orcid":null},"institutions":[{"id":"https://openalex.org/I76130692","display_name":"Zhejiang University","ror":"https://ror.org/00a2xv884","country_code":"CN","type":"education","lineage":["https://openalex.org/I76130692"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Boren Hu","raw_affiliation_strings":["Zhejiang University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Zhejiang University","institution_ids":["https://openalex.org/I76130692"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5054615501","display_name":"Yun Zhu","orcid":"https://orcid.org/0000-0003-3884-4359"},"institutions":[{"id":"https://openalex.org/I76130692","display_name":"Zhejiang University","ror":"https://ror.org/00a2xv884","country_code":"CN","type":"education","lineage":["https://openalex.org/I76130692"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yun Zhu","raw_affiliation_strings":["Zhejiang University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Zhejiang University","institution_ids":["https://openalex.org/I76130692"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100403788","display_name":"Jiacheng Li","orcid":"https://orcid.org/0000-0003-2964-6107"},"institutions":[{"id":"https://openalex.org/I76130692","display_name":"Zhejiang University","ror":"https://ror.org/00a2xv884","country_code":"CN","type":"education","lineage":["https://openalex.org/I76130692"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jiacheng Li","raw_affiliation_strings":["Zhejiang University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Zhejiang University","institution_ids":["https://openalex.org/I76130692"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5063062444","display_name":"Siliang Tang","orcid":"https://orcid.org/0000-0002-7356-9711"},"institutions":[{"id":"https://openalex.org/I76130692","display_name":"Zhejiang University","ror":"https://ror.org/00a2xv884","country_code":"CN","type":"education","lineage":["https://openalex.org/I76130692"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Siliang Tang","raw_affiliation_strings":["Zhejiang University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Zhejiang University","institution_ids":["https://openalex.org/I76130692"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I76130692"],"apc_list":null,"apc_paid":null,"fwci":1.4622,"has_fulltext":false,"cited_by_count":7,"citation_normalized_percentile":{"value":0.84567901,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":89,"max":98},"biblio":{"volume":null,"issue":null,"first_page":"5067","last_page":"5075"},"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":0.9988999962806702,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.9969000220298767,"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":0.8135921955108643},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.7860185503959656},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.7186630368232727},{"id":"https://openalex.org/keywords/benchmark","display_name":"Benchmark (surveying)","score":0.5823078155517578},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.5236149430274963},{"id":"https://openalex.org/keywords/layer","display_name":"Layer (electronics)","score":0.4404257833957672},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.39412856101989746},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.2905253767967224}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8135921955108643},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.7860185503959656},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.7186630368232727},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.5823078155517578},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5236149430274963},{"id":"https://openalex.org/C2779227376","wikidata":"https://www.wikidata.org/wiki/Q6505497","display_name":"Layer (electronics)","level":2,"score":0.4404257833957672},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.39412856101989746},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.2905253767967224},{"id":"https://openalex.org/C185592680","wikidata":"https://www.wikidata.org/wiki/Q2329","display_name":"Chemistry","level":0,"score":0.0},{"id":"https://openalex.org/C205649164","wikidata":"https://www.wikidata.org/wiki/Q1071","display_name":"Geography","level":0,"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/C13280743","wikidata":"https://www.wikidata.org/wiki/Q131089","display_name":"Geodesy","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.24963/ijcai.2023/563","is_oa":true,"landing_page_url":"https://doi.org/10.24963/ijcai.2023/563","pdf_url":"https://www.ijcai.org/proceedings/2023/0563.pdf","source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the Thirty-Second International Joint Conference on Artificial Intelligence","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.24963/ijcai.2023/563","is_oa":true,"landing_page_url":"https://doi.org/10.24963/ijcai.2023/563","pdf_url":"https://www.ijcai.org/proceedings/2023/0563.pdf","source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the Thirty-Second International Joint Conference on Artificial Intelligence","raw_type":"proceedings-article"},"sustainable_development_goals":[{"score":0.46000000834465027,"id":"https://metadata.un.org/sdg/4","display_name":"Quality Education"}],"awards":[{"id":"https://openalex.org/G1781607827","display_name":null,"funder_award_id":"62272411","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"}],"funders":[{"id":"https://openalex.org/F4320306076","display_name":"National Science Foundation","ror":"https://ror.org/021nxhr62"},{"id":"https://openalex.org/F4320318398","display_name":"Ant Group","ror":null},{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"},{"id":"https://openalex.org/F4320322927","display_name":"Zhejiang University","ror":"https://ror.org/00a2xv884"}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4385768057.pdf","grobid_xml":"https://content.openalex.org/works/W4385768057.grobid-xml"},"referenced_works_count":30,"referenced_works":["https://openalex.org/W131533222","https://openalex.org/W1522301498","https://openalex.org/W1599016936","https://openalex.org/W1821462560","https://openalex.org/W2251939518","https://openalex.org/W2923014074","https://openalex.org/W2962965870","https://openalex.org/W2963122961","https://openalex.org/W2963363373","https://openalex.org/W2963393494","https://openalex.org/W2963748441","https://openalex.org/W2963846996","https://openalex.org/W2965373594","https://openalex.org/W2970597249","https://openalex.org/W2978017171","https://openalex.org/W2978670439","https://openalex.org/W2980282514","https://openalex.org/W2996428491","https://openalex.org/W3005680577","https://openalex.org/W3034292689","https://openalex.org/W3035030897","https://openalex.org/W3035038672","https://openalex.org/W3035524453","https://openalex.org/W3038012435","https://openalex.org/W3095602948","https://openalex.org/W3105966348","https://openalex.org/W3156636935","https://openalex.org/W3177265267","https://openalex.org/W4306353146","https://openalex.org/W4394643672"],"related_works":["https://openalex.org/W2378211422","https://openalex.org/W4321353415","https://openalex.org/W2745001401","https://openalex.org/W2130974462","https://openalex.org/W2028665553","https://openalex.org/W2086519370","https://openalex.org/W4246352526","https://openalex.org/W2121910908","https://openalex.org/W915438175","https://openalex.org/W4230315250"],"abstract_inverted_index":{"Dynamic":[0],"early":[1,28,52,114,185],"exiting":[2,29,53,70,186],"has":[3],"been":[4],"proven":[5],"to":[6,87,102],"improve":[7],"the":[8,12,104,118,146,184,193],"inference":[9,60,127],"speed":[10],"of":[11,75,121,145],"pre-trained":[13],"language":[14],"model":[15],"like":[16],"BERT.":[17,76],"However,":[18],"all":[19,24],"samples":[20,33],"must":[21],"go":[22,35],"through":[23,36],"consecutive":[25],"layers":[26,82,106],"before":[27],"and":[30,68,83,95,107,126,165,174,192],"more":[31,37],"complex":[32,179],"usually":[34],"layers,":[38],"which":[39,63,109],"still":[40],"exists":[41],"redundant":[42],"computation.":[43,200],"In":[44],"this":[45],"paper,":[46],"we":[47,90,129,181],"propose":[48,91,130],"a":[49,65,131],"novel":[50],"dynamic":[51],"combined":[54],"with":[55,158,163],"layer":[56,74],"skipping":[57,66,122,194],"for":[58,113,198],"BERT":[59,164],"named":[61],"SmartBERT,":[62],"adds":[64],"gate":[67],"an":[69],"operator":[71],"into":[72,98],"each":[73],"SmartBERT":[77,153],"can":[78],"adaptively":[79,84],"skip":[80],"some":[81,178],"choose":[85],"whether":[86],"exit.":[88],"Besides,":[89],"cross-layer":[92],"contrastive":[93],"learning":[94],"combine":[96],"it":[97],"our":[99,166],"training":[100,125,136],"phases":[101],"boost":[103],"intermediate":[105],"classifiers":[108],"would":[110],"be":[111],"beneficial":[112],"exiting.":[115],"To":[116],"keep":[117],"inconsistent":[119],"usage":[120],"gates":[123],"between":[124],"phases,":[128],"hard":[132],"weight":[133],"mechanism":[134,195],"during":[135],"phase.":[137],"We":[138],"conduct":[139],"experiments":[140],"on":[141,188],"eight":[142],"classification":[143],"datasets":[144],"GLUE":[147],"benchmark.":[148],"Experimental":[149],"results":[150],"show":[151],"that":[152,183],"achieves":[154],"2-3\u00d7":[155],"computation":[156],"reduction":[157],"minimal":[159],"accuracy":[160],"drops":[161],"compared":[162],"method":[167],"outperforms":[168],"previous":[169],"methods":[170],"in":[171,177],"both":[172],"efficiency":[173],"accuracy.":[175],"Moreover,":[176],"datasets,":[180],"prove":[182],"based":[187],"entropy":[189],"hardly":[190],"works,":[191],"is":[196],"essential":[197],"reducing":[199]},"counts_by_year":[{"year":2026,"cited_by_count":2},{"year":2025,"cited_by_count":3},{"year":2024,"cited_by_count":1},{"year":2023,"cited_by_count":1}],"updated_date":"2026-08-18T07:49:30.821534","created_date":"2025-10-10T00:00:00"}