{"id":"https://openalex.org/W7138314058","doi":"https://doi.org/10.1609/aaai.v40i28.39501","title":"Learnable Permutation for Structured Sparsity on Transformer Models","display_name":"Learnable Permutation for Structured Sparsity on Transformer Models","publication_year":2026,"publication_date":"2026-03-14","ids":{"openalex":"https://openalex.org/W7138314058","doi":"https://doi.org/10.1609/aaai.v40i28.39501"},"language":"en","primary_location":{"id":"doi:10.1609/aaai.v40i28.39501","is_oa":true,"landing_page_url":"https://doi.org/10.1609/aaai.v40i28.39501","pdf_url":"https://ojs.aaai.org/index.php/AAAI/article/download/39501/43462","source":{"id":"https://openalex.org/S4210191458","display_name":"Proceedings of the AAAI Conference on Artificial Intelligence","issn_l":"2159-5399","issn":["2159-5399","2374-3468"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/P4310320058","host_organization_name":"Association for the Advancement of Artificial Intelligence","host_organization_lineage":["https://openalex.org/P4310320058"],"host_organization_lineage_names":["Association for the Advancement of Artificial Intelligence"],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the AAAI Conference on Artificial Intelligence","raw_type":"journal-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"diamond","oa_url":"https://ojs.aaai.org/index.php/AAAI/article/download/39501/43462","any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":null,"display_name":"Zekai Li","orcid":null},"institutions":[{"id":"https://openalex.org/I1311921367","display_name":"Advanced Micro Devices (Canada)","ror":"https://ror.org/02yh0k313","country_code":"CA","type":"company","lineage":["https://openalex.org/I1311921367","https://openalex.org/I4210137977"]}],"countries":["CA"],"is_corresponding":false,"raw_author_name":"Zekai Li","raw_affiliation_strings":["Advanced Micro Devices, Inc"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Advanced Micro Devices, Inc","institution_ids":["https://openalex.org/I1311921367"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Ji Liu","orcid":null},"institutions":[{"id":"https://openalex.org/I1311921367","display_name":"Advanced Micro Devices (Canada)","ror":"https://ror.org/02yh0k313","country_code":"CA","type":"company","lineage":["https://openalex.org/I1311921367","https://openalex.org/I4210137977"]}],"countries":["CA"],"is_corresponding":false,"raw_author_name":"Ji Liu","raw_affiliation_strings":["Advanced Micro Devices, Inc"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Advanced Micro Devices, Inc","institution_ids":["https://openalex.org/I1311921367"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Guanchen Li","orcid":null},"institutions":[{"id":"https://openalex.org/I1311921367","display_name":"Advanced Micro Devices (Canada)","ror":"https://ror.org/02yh0k313","country_code":"CA","type":"company","lineage":["https://openalex.org/I1311921367","https://openalex.org/I4210137977"]}],"countries":["CA"],"is_corresponding":false,"raw_author_name":"Guanchen Li","raw_affiliation_strings":["Advanced Micro Devices, Inc"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Advanced Micro Devices, Inc","institution_ids":["https://openalex.org/I1311921367"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Yixing Xu","orcid":null},"institutions":[{"id":"https://openalex.org/I1311921367","display_name":"Advanced Micro Devices (Canada)","ror":"https://ror.org/02yh0k313","country_code":"CA","type":"company","lineage":["https://openalex.org/I1311921367","https://openalex.org/I4210137977"]}],"countries":["CA"],"is_corresponding":false,"raw_author_name":"Yixing Xu","raw_affiliation_strings":["Advanced Micro Devices, Inc"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Advanced Micro Devices, Inc","institution_ids":["https://openalex.org/I1311921367"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Ziqiong Liu","orcid":null},"institutions":[{"id":"https://openalex.org/I1311921367","display_name":"Advanced Micro Devices (Canada)","ror":"https://ror.org/02yh0k313","country_code":"CA","type":"company","lineage":["https://openalex.org/I1311921367","https://openalex.org/I4210137977"]}],"countries":["CA"],"is_corresponding":false,"raw_author_name":"Ziqiong Liu","raw_affiliation_strings":["Advanced Micro Devices, Inc"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Advanced Micro Devices, Inc","institution_ids":["https://openalex.org/I1311921367"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Xuanwu Yin","orcid":null},"institutions":[{"id":"https://openalex.org/I1311921367","display_name":"Advanced Micro Devices (Canada)","ror":"https://ror.org/02yh0k313","country_code":"CA","type":"company","lineage":["https://openalex.org/I1311921367","https://openalex.org/I4210137977"]}],"countries":["CA"],"is_corresponding":false,"raw_author_name":"Xuanwu Yin","raw_affiliation_strings":["Advanced Micro Devices, Inc"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Advanced Micro Devices, Inc","institution_ids":["https://openalex.org/I1311921367"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Dong Li","orcid":null},"institutions":[{"id":"https://openalex.org/I1311921367","display_name":"Advanced Micro Devices (Canada)","ror":"https://ror.org/02yh0k313","country_code":"CA","type":"company","lineage":["https://openalex.org/I1311921367","https://openalex.org/I4210137977"]}],"countries":["CA"],"is_corresponding":false,"raw_author_name":"Dong Li","raw_affiliation_strings":["Advanced Micro Devices, Inc"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Advanced Micro Devices, Inc","institution_ids":["https://openalex.org/I1311921367"]}]},{"author_position":"last","author":{"id":null,"display_name":"Emad Barsoum","orcid":null},"institutions":[{"id":"https://openalex.org/I1311921367","display_name":"Advanced Micro Devices (Canada)","ror":"https://ror.org/02yh0k313","country_code":"CA","type":"company","lineage":["https://openalex.org/I1311921367","https://openalex.org/I4210137977"]}],"countries":["CA"],"is_corresponding":false,"raw_author_name":"Emad Barsoum","raw_affiliation_strings":["Advanced Micro Devices, Inc"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Advanced Micro Devices, Inc","institution_ids":["https://openalex.org/I1311921367"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I1311921367"],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":"40","issue":"28","first_page":"23319","last_page":"23327"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11714","display_name":"Multimodal Machine Learning Applications","score":0.364300012588501,"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"}},"topics":[{"id":"https://openalex.org/T11714","display_name":"Multimodal Machine Learning Applications","score":0.364300012588501,"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"}},{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.17749999463558197,"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"}},{"id":"https://openalex.org/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.1145000010728836,"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/permutation","display_name":"Permutation (music)","score":0.6040999889373779},{"id":"https://openalex.org/keywords/weighting","display_name":"Weighting","score":0.46709999442100525},{"id":"https://openalex.org/keywords/partial-permutation","display_name":"Partial permutation","score":0.46000000834465027},{"id":"https://openalex.org/keywords/solver","display_name":"Solver","score":0.4487000107765198},{"id":"https://openalex.org/keywords/bit-reversal-permutation","display_name":"Bit-reversal permutation","score":0.4323999881744385},{"id":"https://openalex.org/keywords/heuristic","display_name":"Heuristic","score":0.42739999294281006},{"id":"https://openalex.org/keywords/differentiable-function","display_name":"Differentiable function","score":0.3725999891757965},{"id":"https://openalex.org/keywords/transformer","display_name":"Transformer","score":0.3555000126361847},{"id":"https://openalex.org/keywords/permutation-matrix","display_name":"Permutation matrix","score":0.3506999909877777}],"concepts":[{"id":"https://openalex.org/C21308566","wikidata":"https://www.wikidata.org/wiki/Q7169365","display_name":"Permutation (music)","level":2,"score":0.6040999889373779},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.529699981212616},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.4738999903202057},{"id":"https://openalex.org/C183115368","wikidata":"https://www.wikidata.org/wiki/Q856577","display_name":"Weighting","level":2,"score":0.46709999442100525},{"id":"https://openalex.org/C145633696","wikidata":"https://www.wikidata.org/wiki/Q1401935","display_name":"Partial permutation","level":4,"score":0.46000000834465027},{"id":"https://openalex.org/C2778770139","wikidata":"https://www.wikidata.org/wiki/Q1966904","display_name":"Solver","level":2,"score":0.4487000107765198},{"id":"https://openalex.org/C175473596","wikidata":"https://www.wikidata.org/wiki/Q4918660","display_name":"Bit-reversal permutation","level":4,"score":0.4323999881744385},{"id":"https://openalex.org/C173801870","wikidata":"https://www.wikidata.org/wiki/Q201413","display_name":"Heuristic","level":2,"score":0.42739999294281006},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.42500001192092896},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.4221999943256378},{"id":"https://openalex.org/C202615002","wikidata":"https://www.wikidata.org/wiki/Q783507","display_name":"Differentiable function","level":2,"score":0.3725999891757965},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.3555000126361847},{"id":"https://openalex.org/C84140500","wikidata":"https://www.wikidata.org/wiki/Q851512","display_name":"Permutation matrix","level":3,"score":0.3506999909877777},{"id":"https://openalex.org/C151376022","wikidata":"https://www.wikidata.org/wiki/Q168698","display_name":"Exponential function","level":2,"score":0.35010001063346863},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.3481000065803528},{"id":"https://openalex.org/C188198153","wikidata":"https://www.wikidata.org/wiki/Q1613840","display_name":"Limiting","level":2,"score":0.305400013923645},{"id":"https://openalex.org/C106487976","wikidata":"https://www.wikidata.org/wiki/Q685816","display_name":"Matrix (chemical analysis)","level":2,"score":0.302700012922287},{"id":"https://openalex.org/C14036430","wikidata":"https://www.wikidata.org/wiki/Q3736076","display_name":"Function (biology)","level":2,"score":0.3003000020980835},{"id":"https://openalex.org/C39927690","wikidata":"https://www.wikidata.org/wiki/Q11197","display_name":"Logarithm","level":2,"score":0.29910001158714294},{"id":"https://openalex.org/C126980161","wikidata":"https://www.wikidata.org/wiki/Q863783","display_name":"Simulated annealing","level":2,"score":0.29510000348091125},{"id":"https://openalex.org/C147764199","wikidata":"https://www.wikidata.org/wiki/Q6865248","display_name":"Minification","level":2,"score":0.2842000126838684},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.2793999910354614},{"id":"https://openalex.org/C200985842","wikidata":"https://www.wikidata.org/wiki/Q3375503","display_name":"Random permutation","level":3,"score":0.2759000062942505},{"id":"https://openalex.org/C165064840","wikidata":"https://www.wikidata.org/wiki/Q1321061","display_name":"Matching (statistics)","level":2,"score":0.2745000123977661},{"id":"https://openalex.org/C197657726","wikidata":"https://www.wikidata.org/wiki/Q174733","display_name":"Bipartite graph","level":3,"score":0.2703000009059906},{"id":"https://openalex.org/C2779231336","wikidata":"https://www.wikidata.org/wiki/Q7534724","display_name":"Sketch","level":2,"score":0.26829999685287476},{"id":"https://openalex.org/C25818197","wikidata":"https://www.wikidata.org/wiki/Q7255477","display_name":"Pseudorandom permutation","level":4,"score":0.2554999887943268},{"id":"https://openalex.org/C108010975","wikidata":"https://www.wikidata.org/wiki/Q500094","display_name":"Pruning","level":2,"score":0.25529998540878296},{"id":"https://openalex.org/C42058472","wikidata":"https://www.wikidata.org/wiki/Q810214","display_name":"Base (topology)","level":2,"score":0.2540999948978424}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1609/aaai.v40i28.39501","is_oa":true,"landing_page_url":"https://doi.org/10.1609/aaai.v40i28.39501","pdf_url":"https://ojs.aaai.org/index.php/AAAI/article/download/39501/43462","source":{"id":"https://openalex.org/S4210191458","display_name":"Proceedings of the AAAI Conference on Artificial Intelligence","issn_l":"2159-5399","issn":["2159-5399","2374-3468"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/P4310320058","host_organization_name":"Association for the Advancement of Artificial Intelligence","host_organization_lineage":["https://openalex.org/P4310320058"],"host_organization_lineage_names":["Association for the Advancement of Artificial Intelligence"],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the AAAI Conference on Artificial Intelligence","raw_type":"journal-article"},{"id":"pmh:oai:ojs.aaai.org:article/39501","is_oa":false,"landing_page_url":"https://ojs.aaai.org/index.php/AAAI/article/view/39501","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"2159-5399","raw_type":"info:eu-repo/semantics/article"}],"best_oa_location":{"id":"doi:10.1609/aaai.v40i28.39501","is_oa":true,"landing_page_url":"https://doi.org/10.1609/aaai.v40i28.39501","pdf_url":"https://ojs.aaai.org/index.php/AAAI/article/download/39501/43462","source":{"id":"https://openalex.org/S4210191458","display_name":"Proceedings of the AAAI Conference on Artificial Intelligence","issn_l":"2159-5399","issn":["2159-5399","2374-3468"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/P4310320058","host_organization_name":"Association for the Advancement of Artificial Intelligence","host_organization_lineage":["https://openalex.org/P4310320058"],"host_organization_lineage_names":["Association for the Advancement of Artificial Intelligence"],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the AAAI Conference on Artificial Intelligence","raw_type":"journal-article"},"sustainable_development_goals":[{"display_name":"Quality Education","score":0.4349406361579895,"id":"https://metadata.un.org/sdg/4"}],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7138314058.pdf","grobid_xml":"https://content.openalex.org/works/W7138314058.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Structured":[0],"sparsity":[1,131],"has":[2],"emerged":[3],"as":[4],"a":[5,84,93,109,113,126,130],"popular":[6],"model":[7,41],"pruning":[8],"technique,":[9],"widely":[10],"adopted":[11],"in":[12,25],"various":[13],"architectures,":[14],"including":[15],"CNNs,":[16],"Transformer":[17,62],"models,":[18],"and":[19,129,148],"especially":[20],"large":[21],"language":[22,149],"models":[23],"(LLMs)":[24],"recent":[26],"years.":[27],"A":[28],"promising":[29],"direction":[30],"to":[31,47,67,98,118,135],"further":[32],"improve":[33],"post-pruning":[34],"performance":[35],"is":[36],"weight":[37,111],"permutation,":[38],"which":[39],"reorders":[40],"weights":[42],"into":[43],"patterns":[44],"more":[45],"amenable":[46],"pruning.":[48],"However,":[49],"the":[50,54,59,75,100,120,138],"exponential":[51],"growth":[52],"of":[53,61,77,102,108],"permutation":[55,88,95,123,139,157],"search":[56],"space":[57],"with":[58],"scale":[60],"architectures":[63],"forces":[64],"most":[65],"methods":[66],"rely":[68],"on":[69,146],"greedy":[70],"or":[71],"heuristic":[72],"algorithms,":[73],"limiting":[74],"effectiveness":[76],"reordering.":[78],"In":[79],"this":[80],"work,":[81],"we":[82],"propose":[83],"novel":[85],"end-to-end":[86],"learnable":[87,94],"framework.":[89],"Our":[90],"method":[91,154],"introduces":[92],"cost":[96,101,127],"matrix":[97,124],"quantify":[99],"swapping":[103],"any":[104],"two":[105],"input":[106],"channels":[107],"given":[110,125],"matrix,":[112,128],"differentiable":[114],"bipartite":[115],"matching":[116],"solver":[117],"obtain":[119],"optimal":[121],"binary":[122],"optimization":[132],"loss":[133],"function":[134],"directly":[136],"optimize":[137],"operator.":[140],"We":[141],"extensively":[142],"validate":[143],"our":[144,153],"approach":[145],"vision":[147],"Transformers,":[150],"demonstrating":[151],"that":[152],"achieves":[155],"state-of-the-art":[156],"results":[158],"for":[159],"structured":[160],"sparsity.":[161]},"counts_by_year":[],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2026-02-04T00:00:00"}