{"id":"https://openalex.org/W4392181806","doi":"https://doi.org/10.1145/3649329.3658498","title":"APTQ: Attention-aware Post-Training Mixed-Precision Quantization for Large Language Models","display_name":"APTQ: Attention-aware Post-Training Mixed-Precision Quantization for Large Language Models","publication_year":2024,"publication_date":"2024-06-23","ids":{"openalex":"https://openalex.org/W4392181806","doi":"https://doi.org/10.1145/3649329.3658498"},"language":"en","primary_location":{"id":"doi:10.1145/3649329.3658498","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3649329.3658498","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 61st ACM/IEEE Design Automation Conference","raw_type":"proceedings-article"},"type":"preprint","indexed_in":["arxiv","crossref"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2402.14866","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5108870116","display_name":"Ziyi Guan","orcid":null},"institutions":[{"id":"https://openalex.org/I3045169105","display_name":"Southern University of Science and Technology","ror":"https://ror.org/049tv2d57","country_code":"CN","type":"education","lineage":["https://openalex.org/I3045169105"]},{"id":"https://openalex.org/I889458895","display_name":"University of Hong Kong","ror":"https://ror.org/02zhqgq86","country_code":"HK","type":"education","lineage":["https://openalex.org/I889458895"]}],"countries":["CN","HK"],"is_corresponding":true,"raw_author_name":"Ziyi Guan","raw_affiliation_strings":["Department of Electrical and Electronic Engineering, The University of Hong Kong, Hong Kong, Hong Kong","School of Microelectronics, Southern University of Science and Technology, ShenZhen, Guangdong, China"],"raw_orcid":"https://orcid.org/0000-0001-9866-6588","affiliations":[{"raw_affiliation_string":"Department of Electrical and Electronic Engineering, The University of Hong Kong, Hong Kong, Hong Kong","institution_ids":["https://openalex.org/I889458895"]},{"raw_affiliation_string":"School of Microelectronics, Southern University of Science and Technology, ShenZhen, Guangdong, China","institution_ids":["https://openalex.org/I3045169105"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5078800069","display_name":"Hantao Huang","orcid":"https://orcid.org/0000-0002-2533-2082"},"institutions":[{"id":"https://openalex.org/I3045169105","display_name":"Southern University of Science and Technology","ror":"https://ror.org/049tv2d57","country_code":"CN","type":"education","lineage":["https://openalex.org/I3045169105"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Hantao Huang","raw_affiliation_strings":["School of Microelectronics, Southern University of Science and Technology, ShenZhen, Guangdong, China"],"raw_orcid":"https://orcid.org/0000-0002-2533-2082","affiliations":[{"raw_affiliation_string":"School of Microelectronics, Southern University of Science and Technology, ShenZhen, Guangdong, China","institution_ids":["https://openalex.org/I3045169105"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101309752","display_name":"Yupeng Su","orcid":"https://orcid.org/0009-0007-0575-4038"},"institutions":[{"id":"https://openalex.org/I3045169105","display_name":"Southern University of Science and Technology","ror":"https://ror.org/049tv2d57","country_code":"CN","type":"education","lineage":["https://openalex.org/I3045169105"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yupeng Su","raw_affiliation_strings":["School of Microelectronics, Southern University of Science and Technology, ShenZhen, Guangdong, China"],"raw_orcid":"https://orcid.org/0009-0007-0575-4038","affiliations":[{"raw_affiliation_string":"School of Microelectronics, Southern University of Science and Technology, ShenZhen, Guangdong, China","institution_ids":["https://openalex.org/I3045169105"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5108493115","display_name":"Hong Huang","orcid":"https://orcid.org/0009-0002-2448-169X"},"institutions":[{"id":"https://openalex.org/I3045169105","display_name":"Southern University of Science and Technology","ror":"https://ror.org/049tv2d57","country_code":"CN","type":"education","lineage":["https://openalex.org/I3045169105"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Hong Huang","raw_affiliation_strings":["School of Microelectronics, Southern University of Science and Technology, ShenZhen, Guangdong, China"],"raw_orcid":"https://orcid.org/0009-0002-2448-169X","affiliations":[{"raw_affiliation_string":"School of Microelectronics, Southern University of Science and Technology, ShenZhen, Guangdong, China","institution_ids":["https://openalex.org/I3045169105"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5072120057","display_name":"Ngai Wong","orcid":"https://orcid.org/0000-0002-3026-0108"},"institutions":[{"id":"https://openalex.org/I889458895","display_name":"University of Hong Kong","ror":"https://ror.org/02zhqgq86","country_code":"HK","type":"education","lineage":["https://openalex.org/I889458895"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Ngai Wong","raw_affiliation_strings":["Department of Electrical and Electronic Engineering, The University of Hong Kong, Hong Kong, Hong Kong Special Administrative Region of China"],"raw_orcid":"https://orcid.org/0000-0002-3026-0108","affiliations":[{"raw_affiliation_string":"Department of Electrical and Electronic Engineering, The University of Hong Kong, Hong Kong, Hong Kong Special Administrative Region of China","institution_ids":["https://openalex.org/I889458895"]}]},{"author_position":"last","author":{"id":null,"display_name":"Hao Yu","orcid":"https://orcid.org/0000-0002-2674-4118"},"institutions":[{"id":"https://openalex.org/I3045169105","display_name":"Southern University of Science and Technology","ror":"https://ror.org/049tv2d57","country_code":"CN","type":"education","lineage":["https://openalex.org/I3045169105"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Hao Yu","raw_affiliation_strings":["School of Microelectronics, Southern University of Science and Technology, ShenZhen, Guangdong, China"],"raw_orcid":"https://orcid.org/0000-0002-2674-4118","affiliations":[{"raw_affiliation_string":"School of Microelectronics, Southern University of Science and Technology, ShenZhen, Guangdong, China","institution_ids":["https://openalex.org/I3045169105"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":6,"corresponding_author_ids":["https://openalex.org/A5108870116"],"corresponding_institution_ids":["https://openalex.org/I3045169105","https://openalex.org/I889458895"],"apc_list":null,"apc_paid":null,"fwci":6.8334,"has_fulltext":true,"cited_by_count":21,"citation_normalized_percentile":{"value":0.97208885,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":90,"max":100},"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"6"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.9947999715805054,"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":0.9947999715805054,"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":0.9879000186920166,"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/T10201","display_name":"Speech Recognition and Synthesis","score":0.9424999952316284,"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/quantization","display_name":"Quantization (signal processing)","score":0.6406156420707703},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.5560742616653442},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.43604856729507446},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.39154067635536194},{"id":"https://openalex.org/keywords/psychology","display_name":"Psychology","score":0.37814822793006897},{"id":"https://openalex.org/keywords/cognitive-psychology","display_name":"Cognitive psychology","score":0.3496814966201782},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.3392508625984192},{"id":"https://openalex.org/keywords/natural-language-processing","display_name":"Natural language processing","score":0.32404547929763794},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.1932257115840912}],"concepts":[{"id":"https://openalex.org/C28855332","wikidata":"https://www.wikidata.org/wiki/Q198099","display_name":"Quantization (signal processing)","level":2,"score":0.6406156420707703},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5560742616653442},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.43604856729507446},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.39154067635536194},{"id":"https://openalex.org/C15744967","wikidata":"https://www.wikidata.org/wiki/Q9418","display_name":"Psychology","level":0,"score":0.37814822793006897},{"id":"https://openalex.org/C180747234","wikidata":"https://www.wikidata.org/wiki/Q23373","display_name":"Cognitive psychology","level":1,"score":0.3496814966201782},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3392508625984192},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.32404547929763794},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.1932257115840912}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1145/3649329.3658498","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3649329.3658498","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 61st ACM/IEEE Design Automation Conference","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2402.14866","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2402.14866","pdf_url":"https://arxiv.org/pdf/2402.14866","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2402.14866","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2402.14866","pdf_url":"https://arxiv.org/pdf/2402.14866","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[{"id":"https://metadata.un.org/sdg/4","score":0.6700000166893005,"display_name":"Quality Education"}],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":false},"content_urls":{"pdf":"https://content.openalex.org/works/W4392181806.pdf"},"referenced_works_count":6,"referenced_works":["https://openalex.org/W2798603777","https://openalex.org/W2989530497","https://openalex.org/W3095194728","https://openalex.org/W4226278401","https://openalex.org/W4293166090","https://openalex.org/W6969159101"],"related_works":["https://openalex.org/W2169518243","https://openalex.org/W2979160909","https://openalex.org/W2114837856","https://openalex.org/W2961085424","https://openalex.org/W4386322429","https://openalex.org/W2359364609","https://openalex.org/W2322476848","https://openalex.org/W3206280435","https://openalex.org/W2952965081","https://openalex.org/W2099355001"],"abstract_inverted_index":{"Large":[0],"Language":[1],"Models":[2],"(LLMs)":[3],"have":[4],"greatly":[5],"advanced":[6],"the":[7,13,46,56,59,66,71,113],"natural":[8],"language":[9],"processing":[10],"paradigm.":[11],"However,":[12],"high":[14],"computational":[15],"load":[16],"and":[17,125,135],"huge":[18],"model":[19,88],"sizes":[20],"pose":[21],"a":[22,75,104],"grand":[23],"challenge":[24],"for":[25,40,55,78],"deployment":[26],"on":[27,65],"edge":[28],"devices.":[29],"To":[30],"this":[31],"end,":[32],"we":[33],"propose":[34],"APTQ":[35,92,118],"(Attention-aware":[36],"Post-Training":[37],"Mixed-Precision":[38],"Quantization)":[39],"LLMs,":[41],"which":[42],"considers":[43],"not":[44],"only":[45],"second-order":[47],"information":[48],"of":[49,62,100,123,131],"each":[50],"layer's":[51],"weights,":[52],"but":[53],"also,":[54],"first":[57],"time,":[58],"nonlinear":[60],"effect":[61],"attention":[63],"outputs":[64],"entire":[67],"model.":[68],"We":[69],"leverage":[70],"Hessian":[72],"trace":[73],"as":[74],"sensitivity":[76],"metric":[77],"mixed-precision":[79],"quantization,":[80],"ensuring":[81],"an":[82,98,128],"informed":[83],"precision":[84,111],"reduction":[85],"that":[86],"retains":[87],"performance.":[89],"Experiments":[90],"show":[91],"surpasses":[93],"previous":[94],"quantization":[95],"methods,":[96],"achieving":[97],"average":[99,129],"4":[101],"bit":[102],"width":[103],"5.22":[105],"perplexity":[106],"nearly":[107],"equivalent":[108],"to":[109,141],"full":[110],"in":[112,133],"C4":[114],"dataset.":[115],"In":[116],"addition,":[117],"attains":[119],"state-of-the-art":[120],"zero-shot":[121],"accuracy":[122],"68.24%":[124],"70.48%":[126],"at":[127],"bitwidth":[130],"3.8":[132],"LLaMa-7B":[134],"LLaMa-13B,":[136],"respectively,":[137],"demonstrating":[138],"its":[139],"effectiveness":[140],"produce":[142],"high-quality":[143],"quantized":[144],"LLMs.":[145]},"counts_by_year":[{"year":2026,"cited_by_count":3},{"year":2025,"cited_by_count":17},{"year":2024,"cited_by_count":1}],"updated_date":"2026-05-06T08:25:59.206177","created_date":"2024-02-27T00:00:00"}