{"id":"https://openalex.org/W4402982216","doi":"https://doi.org/10.1109/icme57554.2024.10688089","title":"Integer or Floating Point? New Outlooks for Low-Bit Quantization on Large Language Models","display_name":"Integer or Floating Point? New Outlooks for Low-Bit Quantization on Large Language Models","publication_year":2024,"publication_date":"2024-07-15","ids":{"openalex":"https://openalex.org/W4402982216","doi":"https://doi.org/10.1109/icme57554.2024.10688089"},"language":"en","primary_location":{"id":"doi:10.1109/icme57554.2024.10688089","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icme57554.2024.10688089","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2024 IEEE International Conference on Multimedia and Expo (ICME)","raw_type":"proceedings-article"},"type":"conference-paper","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/A5046020427","display_name":"Yijia Zhang","orcid":"https://orcid.org/0000-0002-5843-4675"},"institutions":[{"id":"https://openalex.org/I183067930","display_name":"Shanghai Jiao Tong University","ror":"https://ror.org/0220qvk04","country_code":"CN","type":"education","lineage":["https://openalex.org/I183067930"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yijia Zhang","raw_affiliation_strings":["Shanghai Jiao Tong University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Shanghai Jiao Tong University","institution_ids":["https://openalex.org/I183067930"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102973561","display_name":"Lingran Zhao","orcid":"https://orcid.org/0009-0007-1313-789X"},"institutions":[{"id":"https://openalex.org/I20231570","display_name":"Peking University","ror":"https://ror.org/02v51f717","country_code":"CN","type":"education","lineage":["https://openalex.org/I20231570"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Lingran Zhao","raw_affiliation_strings":["Peking University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Peking University","institution_ids":["https://openalex.org/I20231570"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101401413","display_name":"Shijie Cao","orcid":"https://orcid.org/0000-0002-5646-2947"},"institutions":[{"id":"https://openalex.org/I4210113369","display_name":"Microsoft Research Asia (China)","ror":"https://ror.org/0300m5276","country_code":"CN","type":"company","lineage":["https://openalex.org/I1290206253","https://openalex.org/I4210113369"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Shijie Cao","raw_affiliation_strings":["Microsoft Research Asia"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Microsoft Research Asia","institution_ids":["https://openalex.org/I4210113369"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101606837","display_name":"Sicheng Zhang","orcid":"https://orcid.org/0009-0000-0128-5472"},"institutions":[{"id":"https://openalex.org/I183067930","display_name":"Shanghai Jiao Tong University","ror":"https://ror.org/0220qvk04","country_code":"CN","type":"education","lineage":["https://openalex.org/I183067930"]},{"id":"https://openalex.org/I20231570","display_name":"Peking University","ror":"https://ror.org/02v51f717","country_code":"CN","type":"education","lineage":["https://openalex.org/I20231570"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Sicheng Zhang","raw_affiliation_strings":["Peking University","Shanghai Jiao Tong University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Peking University","institution_ids":["https://openalex.org/I20231570"]},{"raw_affiliation_string":"Shanghai Jiao Tong University","institution_ids":["https://openalex.org/I183067930"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100628303","display_name":"Wenqiang Wang","orcid":"https://orcid.org/0000-0002-4425-6636"},"institutions":[{"id":"https://openalex.org/I183067930","display_name":"Shanghai Jiao Tong University","ror":"https://ror.org/0220qvk04","country_code":"CN","type":"education","lineage":["https://openalex.org/I183067930"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Wenqiang Wang","raw_affiliation_strings":["Shanghai Jiao Tong University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Shanghai Jiao Tong University","institution_ids":["https://openalex.org/I183067930"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5103092945","display_name":"Ting Cao","orcid":null},"institutions":[{"id":"https://openalex.org/I4210113369","display_name":"Microsoft Research Asia (China)","ror":"https://ror.org/0300m5276","country_code":"CN","type":"company","lineage":["https://openalex.org/I1290206253","https://openalex.org/I4210113369"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Ting Cao","raw_affiliation_strings":["Microsoft Research Asia"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Microsoft Research Asia","institution_ids":["https://openalex.org/I4210113369"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5002213259","display_name":"Fan Yang","orcid":"https://orcid.org/0000-0002-6039-8212"},"institutions":[{"id":"https://openalex.org/I4210113369","display_name":"Microsoft Research Asia (China)","ror":"https://ror.org/0300m5276","country_code":"CN","type":"company","lineage":["https://openalex.org/I1290206253","https://openalex.org/I4210113369"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Fan Yang","raw_affiliation_strings":["Microsoft Research Asia"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Microsoft Research Asia","institution_ids":["https://openalex.org/I4210113369"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5103009688","display_name":"Mao Yang","orcid":"https://orcid.org/0000-0003-1488-4671"},"institutions":[{"id":"https://openalex.org/I4210113369","display_name":"Microsoft Research Asia (China)","ror":"https://ror.org/0300m5276","country_code":"CN","type":"company","lineage":["https://openalex.org/I1290206253","https://openalex.org/I4210113369"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Mao Yang","raw_affiliation_strings":["Microsoft Research Asia"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Microsoft Research Asia","institution_ids":["https://openalex.org/I4210113369"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5013030532","display_name":"Shanghang Zhang","orcid":"https://orcid.org/0000-0003-4047-3526"},"institutions":[{"id":"https://openalex.org/I183067930","display_name":"Shanghai Jiao Tong University","ror":"https://ror.org/0220qvk04","country_code":"CN","type":"education","lineage":["https://openalex.org/I183067930"]},{"id":"https://openalex.org/I20231570","display_name":"Peking University","ror":"https://ror.org/02v51f717","country_code":"CN","type":"education","lineage":["https://openalex.org/I20231570"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Shanghang Zhang","raw_affiliation_strings":["Peking University","Shanghai Jiao Tong University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Peking University","institution_ids":["https://openalex.org/I20231570"]},{"raw_affiliation_string":"Shanghai Jiao Tong University","institution_ids":["https://openalex.org/I183067930"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5101205509","display_name":"Ningyi Xu","orcid":null},"institutions":[{"id":"https://openalex.org/I183067930","display_name":"Shanghai Jiao Tong University","ror":"https://ror.org/0220qvk04","country_code":"CN","type":"education","lineage":["https://openalex.org/I183067930"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Ningyi Xu","raw_affiliation_strings":["Shanghai Jiao Tong University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Shanghai Jiao Tong University","institution_ids":["https://openalex.org/I183067930"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":3,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":7,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"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":0.6984999775886536,"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":0.6984999775886536,"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/T10320","display_name":"Neural Networks and Applications","score":0.635200023651123,"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":0.7194200158119202},{"id":"https://openalex.org/keywords/floating-point","display_name":"Floating point","score":0.6197241544723511},{"id":"https://openalex.org/keywords/integer","display_name":"Integer (computer science)","score":0.6087478399276733},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.5537006855010986},{"id":"https://openalex.org/keywords/bit","display_name":"Bit (key)","score":0.546185314655304},{"id":"https://openalex.org/keywords/arithmetic","display_name":"Arithmetic","score":0.4944288730621338},{"id":"https://openalex.org/keywords/point","display_name":"Point (geometry)","score":0.47833701968193054},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.2549855411052704},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.2326573133468628},{"id":"https://openalex.org/keywords/programming-language","display_name":"Programming language","score":0.21101626753807068},{"id":"https://openalex.org/keywords/computer-network","display_name":"Computer network","score":0.1023288369178772}],"concepts":[{"id":"https://openalex.org/C28855332","wikidata":"https://www.wikidata.org/wiki/Q198099","display_name":"Quantization (signal processing)","level":2,"score":0.7194200158119202},{"id":"https://openalex.org/C84211073","wikidata":"https://www.wikidata.org/wiki/Q117879","display_name":"Floating point","level":2,"score":0.6197241544723511},{"id":"https://openalex.org/C97137487","wikidata":"https://www.wikidata.org/wiki/Q729138","display_name":"Integer (computer science)","level":2,"score":0.6087478399276733},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5537006855010986},{"id":"https://openalex.org/C117011727","wikidata":"https://www.wikidata.org/wiki/Q1278488","display_name":"Bit (key)","level":2,"score":0.546185314655304},{"id":"https://openalex.org/C94375191","wikidata":"https://www.wikidata.org/wiki/Q11205","display_name":"Arithmetic","level":1,"score":0.4944288730621338},{"id":"https://openalex.org/C28719098","wikidata":"https://www.wikidata.org/wiki/Q44946","display_name":"Point (geometry)","level":2,"score":0.47833701968193054},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.2549855411052704},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.2326573133468628},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.21101626753807068},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.1023288369178772},{"id":"https://openalex.org/C2524010","wikidata":"https://www.wikidata.org/wiki/Q8087","display_name":"Geometry","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/icme57554.2024.10688089","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icme57554.2024.10688089","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2024 IEEE International Conference on Multimedia and Expo (ICME)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":8,"referenced_works":["https://openalex.org/W2946609015","https://openalex.org/W2963015836","https://openalex.org/W2963122961","https://openalex.org/W2998617917","https://openalex.org/W3017024317","https://openalex.org/W3159727696","https://openalex.org/W3172967059","https://openalex.org/W4386160289"],"related_works":["https://openalex.org/W2411923897","https://openalex.org/W4394546135","https://openalex.org/W4255129386","https://openalex.org/W4245068234","https://openalex.org/W2025161972","https://openalex.org/W1978352992","https://openalex.org/W1585685783","https://openalex.org/W2119414612","https://openalex.org/W2268486090","https://openalex.org/W1988006935"],"abstract_inverted_index":{"Efficient":[0],"deployment":[1],"of":[2,81],"Large":[3],"Language":[4],"Models":[5],"(LLMs)":[6],"requires":[7],"low-bit":[8,18,29],"quantization":[9,26,62,104],"to":[10],"reduce":[11],"model":[12],"size":[13],"and":[14,100],"inference":[15],"cost.":[16],"Besides":[17],"integer":[19],"formats":[20,31,51],"(e.g.,":[21,32],"INT8/INT4)":[22],"used":[23],"in":[24,97],"previous":[25],"works,":[27],"emerging":[28],"floating-point":[30,50],"FP8/FP4)":[33],"supported":[34],"by":[35],"advanced":[36],"hardware":[37,112],"like":[38],"NVIDIA\u2019s":[39],"H100":[40],"GPU":[41],"offer":[42],"an":[43],"alternative.":[44],"Our":[45,86],"study":[46],"finds":[47],"that":[48,59],"introducing":[49],"significantly":[52],"improves":[53],"LLMs":[54],"quantization.":[55],"We":[56],"also":[57],"discover":[58],"the":[60,70,79],"optimal":[61,71],"format":[63,72],"varies":[64],"across":[65,106],"layers.":[66],"Therefore,":[67],"we":[68,77],"select":[69],"for":[73],"each":[74],"layer,":[75],"which":[76],"call":[78],"Mixture":[80],"Formats":[82],"Quantization":[83],"(MoFQ)":[84],"method.":[85],"MoFQ":[87],"method":[88],"achieves":[89],"better":[90],"or":[91],"comparable":[92],"results":[93],"over":[94],"current":[95],"methods":[96],"weight-only":[98],"(W-only)":[99],"weight-activation":[101],"(WA)":[102],"post-training":[103],"scenarios":[105],"various":[107],"tasks,":[108],"with":[109],"no":[110],"additional":[111],"overhead.":[113]},"counts_by_year":[{"year":2025,"cited_by_count":5},{"year":2024,"cited_by_count":2}],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2025-10-10T00:00:00"}
