{"id":"https://openalex.org/W4225401825","doi":"https://doi.org/10.24963/ijcai.2022/586","title":"Parameter-Efficient Sparsity for Large Language Models Fine-Tuning","display_name":"Parameter-Efficient Sparsity for Large Language Models Fine-Tuning","publication_year":2022,"publication_date":"2022-07-01","ids":{"openalex":"https://openalex.org/W4225401825","doi":"https://doi.org/10.24963/ijcai.2022/586"},"language":"en","primary_location":{"id":"doi:10.24963/ijcai.2022/586","is_oa":true,"landing_page_url":"https://doi.org/10.24963/ijcai.2022/586","pdf_url":"https://www.ijcai.org/proceedings/2022/0586.pdf","source":{"id":"https://openalex.org/S4363608755","display_name":"Proceedings of the Thirty-First International Joint Conference on Artificial Intelligence","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the Thirty-First International Joint Conference on Artificial Intelligence","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"bronze","oa_url":"https://www.ijcai.org/proceedings/2022/0586.pdf","any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5100743492","display_name":"Yuchao Li","orcid":"https://orcid.org/0000-0002-5723-8357"},"institutions":[{"id":"https://openalex.org/I4210095624","display_name":"Alibaba Group (United States)","ror":"https://ror.org/00rn0m335","country_code":"US","type":"company","lineage":["https://openalex.org/I4210095624","https://openalex.org/I45928872"]}],"countries":["US"],"is_corresponding":true,"raw_author_name":"Yuchao Li","raw_affiliation_strings":["Alibaba Group"],"affiliations":[{"raw_affiliation_string":"Alibaba Group","institution_ids":["https://openalex.org/I4210095624"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5019525050","display_name":"Fuli Luo","orcid":"https://orcid.org/0000-0002-5403-6434"},"institutions":[{"id":"https://openalex.org/I4210095624","display_name":"Alibaba Group (United States)","ror":"https://ror.org/00rn0m335","country_code":"US","type":"company","lineage":["https://openalex.org/I4210095624","https://openalex.org/I45928872"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Fuli Luo","raw_affiliation_strings":["Alibaba Group"],"affiliations":[{"raw_affiliation_string":"Alibaba Group","institution_ids":["https://openalex.org/I4210095624"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5065907658","display_name":"Chuanqi Tan","orcid":"https://orcid.org/0000-0002-6676-3057"},"institutions":[{"id":"https://openalex.org/I4210095624","display_name":"Alibaba Group (United States)","ror":"https://ror.org/00rn0m335","country_code":"US","type":"company","lineage":["https://openalex.org/I4210095624","https://openalex.org/I45928872"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Chuanqi Tan","raw_affiliation_strings":["Alibaba Group"],"affiliations":[{"raw_affiliation_string":"Alibaba Group","institution_ids":["https://openalex.org/I4210095624"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100707460","display_name":"Mengdi Wang","orcid":"https://orcid.org/0000-0002-2101-9507"},"institutions":[{"id":"https://openalex.org/I4210095624","display_name":"Alibaba Group (United States)","ror":"https://ror.org/00rn0m335","country_code":"US","type":"company","lineage":["https://openalex.org/I4210095624","https://openalex.org/I45928872"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Mengdi Wang","raw_affiliation_strings":["Alibaba Group"],"affiliations":[{"raw_affiliation_string":"Alibaba Group","institution_ids":["https://openalex.org/I4210095624"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5047856952","display_name":"Songfang Huang","orcid":"https://orcid.org/0000-0001-8084-0904"},"institutions":[{"id":"https://openalex.org/I4210095624","display_name":"Alibaba Group (United States)","ror":"https://ror.org/00rn0m335","country_code":"US","type":"company","lineage":["https://openalex.org/I4210095624","https://openalex.org/I45928872"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Songfang Huang","raw_affiliation_strings":["Alibaba Group"],"affiliations":[{"raw_affiliation_string":"Alibaba Group","institution_ids":["https://openalex.org/I4210095624"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5104910007","display_name":"Shen Li","orcid":"https://orcid.org/0000-0003-2947-3787"},"institutions":[{"id":"https://openalex.org/I4210095624","display_name":"Alibaba Group (United States)","ror":"https://ror.org/00rn0m335","country_code":"US","type":"company","lineage":["https://openalex.org/I4210095624","https://openalex.org/I45928872"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Shen Li","raw_affiliation_strings":["Alibaba Group"],"affiliations":[{"raw_affiliation_string":"Alibaba Group","institution_ids":["https://openalex.org/I4210095624"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5103029322","display_name":"Junjie Bai","orcid":"https://orcid.org/0009-0007-2616-2015"},"institutions":[{"id":"https://openalex.org/I4210095624","display_name":"Alibaba Group (United States)","ror":"https://ror.org/00rn0m335","country_code":"US","type":"company","lineage":["https://openalex.org/I4210095624","https://openalex.org/I45928872"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Junjie Bai","raw_affiliation_strings":["Alibaba Group"],"affiliations":[{"raw_affiliation_string":"Alibaba Group","institution_ids":["https://openalex.org/I4210095624"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":7,"corresponding_author_ids":["https://openalex.org/A5100743492"],"corresponding_institution_ids":["https://openalex.org/I4210095624"],"apc_list":null,"apc_paid":null,"fwci":0.9415,"has_fulltext":false,"cited_by_count":9,"citation_normalized_percentile":{"value":0.75867918,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":90,"max":99},"biblio":{"volume":null,"issue":null,"first_page":"4223","last_page":"4229"},"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":0.9990000128746033,"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":0.9955000281333923,"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/computer-science","display_name":"Computer science","score":0.8132628202438354},{"id":"https://openalex.org/keywords/redundancy","display_name":"Redundancy (engineering)","score":0.6854822635650635},{"id":"https://openalex.org/keywords/sparse-matrix","display_name":"Sparse matrix","score":0.6175297498703003},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.5977268815040588},{"id":"https://openalex.org/keywords/overhead","display_name":"Overhead (engineering)","score":0.5414469838142395},{"id":"https://openalex.org/keywords/memory-footprint","display_name":"Memory footprint","score":0.537096381187439},{"id":"https://openalex.org/keywords/focus","display_name":"Focus (optics)","score":0.5066207051277161},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.42271292209625244},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.4226205348968506},{"id":"https://openalex.org/keywords/training-set","display_name":"Training set","score":0.42196419835090637},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.3983479142189026}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8132628202438354},{"id":"https://openalex.org/C152124472","wikidata":"https://www.wikidata.org/wiki/Q1204361","display_name":"Redundancy (engineering)","level":2,"score":0.6854822635650635},{"id":"https://openalex.org/C56372850","wikidata":"https://www.wikidata.org/wiki/Q1050404","display_name":"Sparse matrix","level":3,"score":0.6175297498703003},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.5977268815040588},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.5414469838142395},{"id":"https://openalex.org/C74912251","wikidata":"https://www.wikidata.org/wiki/Q6815727","display_name":"Memory footprint","level":2,"score":0.537096381187439},{"id":"https://openalex.org/C192209626","wikidata":"https://www.wikidata.org/wiki/Q190909","display_name":"Focus (optics)","level":2,"score":0.5066207051277161},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.42271292209625244},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4226205348968506},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.42196419835090637},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3983479142189026},{"id":"https://openalex.org/C120665830","wikidata":"https://www.wikidata.org/wiki/Q14620","display_name":"Optics","level":1,"score":0.0},{"id":"https://openalex.org/C163716315","wikidata":"https://www.wikidata.org/wiki/Q901177","display_name":"Gaussian","level":2,"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/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.0},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.24963/ijcai.2022/586","is_oa":true,"landing_page_url":"https://doi.org/10.24963/ijcai.2022/586","pdf_url":"https://www.ijcai.org/proceedings/2022/0586.pdf","source":{"id":"https://openalex.org/S4363608755","display_name":"Proceedings of the Thirty-First International Joint Conference on Artificial Intelligence","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the Thirty-First International Joint Conference on Artificial Intelligence","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.24963/ijcai.2022/586","is_oa":true,"landing_page_url":"https://doi.org/10.24963/ijcai.2022/586","pdf_url":"https://www.ijcai.org/proceedings/2022/0586.pdf","source":{"id":"https://openalex.org/S4363608755","display_name":"Proceedings of the Thirty-First International Joint Conference on Artificial Intelligence","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the Thirty-First International Joint Conference on Artificial Intelligence","raw_type":"proceedings-article"},"sustainable_development_goals":[{"display_name":"Quality Education","id":"https://metadata.un.org/sdg/4","score":0.7699999809265137}],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4225401825.pdf","grobid_xml":"https://content.openalex.org/works/W4225401825.grobid-xml"},"referenced_works_count":22,"referenced_works":["https://openalex.org/W2764043458","https://openalex.org/W2896457183","https://openalex.org/W2923014074","https://openalex.org/W2949960976","https://openalex.org/W2964303773","https://openalex.org/W2965373594","https://openalex.org/W2965862774","https://openalex.org/W3005842225","https://openalex.org/W3015982254","https://openalex.org/W3034340181","https://openalex.org/W3038012435","https://openalex.org/W3104216863","https://openalex.org/W3104688113","https://openalex.org/W3168867926","https://openalex.org/W3173390350","https://openalex.org/W3174702398","https://openalex.org/W3205949070","https://openalex.org/W4205991051","https://openalex.org/W4287777801","https://openalex.org/W4288480287","https://openalex.org/W4297813615","https://openalex.org/W4386453467"],"related_works":["https://openalex.org/W2494338568","https://openalex.org/W1495042958","https://openalex.org/W2122678784","https://openalex.org/W2282510344","https://openalex.org/W139987158","https://openalex.org/W2183994405","https://openalex.org/W1875646599","https://openalex.org/W4287644835","https://openalex.org/W3092281475","https://openalex.org/W3098003361"],"abstract_inverted_index":{"With":[0],"the":[1,21,36,39,46,74,89,99,106,134,142,151],"dramatically":[2],"increased":[3],"number":[4,75,186],"of":[5,38,52,76,101,109,127,137,140,168,187],"parameters":[6,78,202],"in":[7,45,82],"language":[8,58],"models,":[9],"sparsity":[10,179,194],"methods":[11],"have":[12],"received":[13],"ever-increasing":[14],"research":[15,25],"focus":[16],"to":[17,29,72,94,132,203],"compress":[18],"and":[19,49,91,96,113,120,155,164],"accelerate":[20],"models.":[22,59],"While":[23],"most":[24],"focuses":[26],"on":[27,123,166,173,207],"how":[28],"accurately":[30,97],"retain":[31],"appropriate":[32],"weights":[33],"while":[34],"maintaining":[35],"performance":[37,206],"compressed":[40],"model,":[41],"there":[42],"are":[43,130],"challenges":[44],"computational":[47],"overhead":[48],"memory":[50],"footprint":[51],"sparse":[53,152],"training":[54,81,153,183],"when":[55],"compressing":[56],"large-scale":[57],"To":[60],"address":[61],"this":[62],"problem,":[63],"we":[64,86,104],"propose":[65],"a":[66,184],"Parameter-efficient":[67],"Sparse":[68],"Training":[69],"(PST)":[70],"method":[71],"reduce":[73],"trainable":[77,201],"during":[79],"sparse-aware":[80],"downstream":[83],"tasks.":[84],"Specifically,":[85],"first":[87],"combine":[88],"data-free":[90],"data-driven":[92,110,135],"criteria":[93],"efficiently":[95],"measure":[98],"importance":[100,112,136,145],"weights.":[102],"Then":[103],"investigate":[105],"intrinsic":[107],"redundancy":[108],"weight":[111],"derive":[114],"two":[115,125],"obvious":[116],"characteristics":[117],"i.e.":[118],"low-rankness":[119],"structuredness.":[121],"Based":[122],"that,":[124],"groups":[126],"small":[128,185],"matrices":[129],"introduced":[131],"compute":[133],"weights,":[138],"instead":[139],"using":[141],"original":[143],"large":[144],"score":[146],"matrix,":[147],"which":[148],"therefore":[149],"makes":[150],"resource-efficient":[154],"parameter-efficient.":[156],"Experiments":[157],"with":[158,192],"diverse":[159],"networks":[160],"(i.e.":[161],"BERT,":[162],"RoBERTa":[163],"GPT-2)":[165],"dozens":[167],"datasets":[169],"demonstrate":[170],"PST":[171,197],"performs":[172],"par":[174],"or":[175],"better":[176],"than":[177],"previous":[178,193],"methods,":[180,195],"despite":[181],"only":[182,198],"parameters.":[188],"For":[189],"instance,":[190],"compared":[191],"our":[196],"requires":[199],"1.5%":[200],"achieve":[204],"comparable":[205],"BERT.":[208]},"counts_by_year":[{"year":2025,"cited_by_count":6},{"year":2024,"cited_by_count":2},{"year":2023,"cited_by_count":1}],"updated_date":"2026-03-27T14:29:43.386196","created_date":"2025-10-10T00:00:00"}