{"id":"https://openalex.org/W4415427983","doi":"https://doi.org/10.3233/faia251342","title":"SDQ-LLM: Sigma-Delta Quantization for 1-Bit LLMs of Any Size","display_name":"SDQ-LLM: Sigma-Delta Quantization for 1-Bit LLMs of Any Size","publication_year":2025,"publication_date":"2025-10-21","ids":{"openalex":"https://openalex.org/W4415427983","doi":"https://doi.org/10.3233/faia251342"},"language":null,"primary_location":{"id":"doi:10.3233/faia251342","is_oa":true,"landing_page_url":"https://doi.org/10.3233/faia251342","pdf_url":null,"source":{"id":"https://openalex.org/S4210201731","display_name":"Frontiers in artificial intelligence and applications","issn_l":"0922-6389","issn":["0922-6389","1879-8314"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"journal"},"license":"cc-by-nc","license_id":"https://openalex.org/licenses/cc-by-nc","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Frontiers in Artificial Intelligence and Applications","raw_type":"book-chapter"},"type":"book-chapter","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"hybrid","oa_url":"https://doi.org/10.3233/faia251342","any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5038194236","display_name":"Jinshi Xia","orcid":null},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":true,"raw_author_name":"Junhao Xia","raw_affiliation_strings":["Dept of Electronic Engineering, Tsinghua University, Beijing, China"],"affiliations":[{"raw_affiliation_string":"Dept of Electronic Engineering, Tsinghua University, Beijing, China","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5113559722","display_name":"Ming Zhao","orcid":"https://orcid.org/0009-0005-8599-9865"},"institutions":[{"id":"https://openalex.org/I4210156423","display_name":"National Engineering Research Center for Information Technology in Agriculture","ror":"https://ror.org/04c3j3t84","country_code":"CN","type":"facility","lineage":["https://openalex.org/I4210156423"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Ming Zhao","raw_affiliation_strings":["Beijing National Research Center for Information Science and Technology, Beijing, China"],"affiliations":[{"raw_affiliation_string":"Beijing National Research Center for Information Science and Technology, Beijing, China","institution_ids":["https://openalex.org/I4210156423"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5069732463","display_name":"Limin Xiao","orcid":"https://orcid.org/0000-0002-8791-3456"},"institutions":[{"id":"https://openalex.org/I4210156423","display_name":"National Engineering Research Center for Information Technology in Agriculture","ror":"https://ror.org/04c3j3t84","country_code":"CN","type":"facility","lineage":["https://openalex.org/I4210156423"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Limin Xiao","raw_affiliation_strings":["Beijing National Research Center for Information Science and Technology, Beijing, China"],"affiliations":[{"raw_affiliation_string":"Beijing National Research Center for Information Science and Technology, Beijing, China","institution_ids":["https://openalex.org/I4210156423"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5100720560","display_name":"Xiujun Zhang","orcid":"https://orcid.org/0000-0001-9402-2777"},"institutions":[{"id":"https://openalex.org/I4210156423","display_name":"National Engineering Research Center for Information Technology in Agriculture","ror":"https://ror.org/04c3j3t84","country_code":"CN","type":"facility","lineage":["https://openalex.org/I4210156423"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xiujun Zhang","raw_affiliation_strings":["Beijing National Research Center for Information Science and Technology, Beijing, China"],"affiliations":[{"raw_affiliation_string":"Beijing National Research Center for Information Science and Technology, Beijing, China","institution_ids":["https://openalex.org/I4210156423"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":4,"corresponding_author_ids":["https://openalex.org/A5038194236"],"corresponding_institution_ids":["https://openalex.org/I99065089"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.6768595,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11181","display_name":"Advanced Data Storage Technologies","score":0.8931999802589417,"subfield":{"id":"https://openalex.org/subfields/1705","display_name":"Computer Networks and Communications"},"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/T11181","display_name":"Advanced Data Storage Technologies","score":0.8931999802589417,"subfield":{"id":"https://openalex.org/subfields/1705","display_name":"Computer Networks and Communications"},"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/T11697","display_name":"Numerical Methods and Algorithms","score":0.8873999714851379,"subfield":{"id":"https://openalex.org/subfields/1703","display_name":"Computational Theory and Mathematics"},"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/T10558","display_name":"Advancements in Semiconductor Devices and Circuit Design","score":0.8596000075340271,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/quantization","display_name":"Quantization (signal processing)","score":0.7516000270843506},{"id":"https://openalex.org/keywords/vector-quantization","display_name":"Vector quantization","score":0.45649999380111694},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.3862999975681305},{"id":"https://openalex.org/keywords/robustness","display_name":"Robustness (evolution)","score":0.3806000053882599},{"id":"https://openalex.org/keywords/smoothing","display_name":"Smoothing","score":0.32260000705718994},{"id":"https://openalex.org/keywords/leverage","display_name":"Leverage (statistics)","score":0.32019999623298645}],"concepts":[{"id":"https://openalex.org/C28855332","wikidata":"https://www.wikidata.org/wiki/Q198099","display_name":"Quantization (signal processing)","level":2,"score":0.7516000270843506},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.49720001220703125},{"id":"https://openalex.org/C199833920","wikidata":"https://www.wikidata.org/wiki/Q612536","display_name":"Vector quantization","level":2,"score":0.45649999380111694},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.4401000142097473},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.4239000082015991},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.3862999975681305},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.3806000053882599},{"id":"https://openalex.org/C3770464","wikidata":"https://www.wikidata.org/wiki/Q775963","display_name":"Smoothing","level":2,"score":0.32260000705718994},{"id":"https://openalex.org/C153083717","wikidata":"https://www.wikidata.org/wiki/Q6535263","display_name":"Leverage (statistics)","level":2,"score":0.32019999623298645},{"id":"https://openalex.org/C150921843","wikidata":"https://www.wikidata.org/wiki/Q1170431","display_name":"Resampling","level":2,"score":0.31459999084472656},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.29989999532699585},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.28459998965263367},{"id":"https://openalex.org/C2781067378","wikidata":"https://www.wikidata.org/wiki/Q17027399","display_name":"Interpretability","level":2,"score":0.2816999852657318},{"id":"https://openalex.org/C47446073","wikidata":"https://www.wikidata.org/wiki/Q5165890","display_name":"Control theory (sociology)","level":3,"score":0.2768000066280365},{"id":"https://openalex.org/C94375191","wikidata":"https://www.wikidata.org/wiki/Q11205","display_name":"Arithmetic","level":1,"score":0.2538999915122986},{"id":"https://openalex.org/C2777413886","wikidata":"https://www.wikidata.org/wiki/Q3276013","display_name":"Fluency","level":2,"score":0.25200000405311584}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.3233/faia251342","is_oa":true,"landing_page_url":"https://doi.org/10.3233/faia251342","pdf_url":null,"source":{"id":"https://openalex.org/S4210201731","display_name":"Frontiers in artificial intelligence and applications","issn_l":"0922-6389","issn":["0922-6389","1879-8314"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"journal"},"license":"cc-by-nc","license_id":"https://openalex.org/licenses/cc-by-nc","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Frontiers in Artificial Intelligence and Applications","raw_type":"book-chapter"}],"best_oa_location":{"id":"doi:10.3233/faia251342","is_oa":true,"landing_page_url":"https://doi.org/10.3233/faia251342","pdf_url":null,"source":{"id":"https://openalex.org/S4210201731","display_name":"Frontiers in artificial intelligence and applications","issn_l":"0922-6389","issn":["0922-6389","1879-8314"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"journal"},"license":"cc-by-nc","license_id":"https://openalex.org/licenses/cc-by-nc","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Frontiers in Artificial Intelligence and Applications","raw_type":"book-chapter"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Large":[0],"language":[1],"models":[2],"(LLMs)":[3],"face":[4],"significant":[5],"computational":[6],"and":[7,84,145,160,170,178,191,199,207,218],"memory":[8,67],"challenges,":[9],"making":[10],"extremely":[11,38,123],"low-bit":[12,39,124],"quantization":[13,40,132,168],"crucial":[14],"for":[15,27,77],"their":[16,45],"efficient":[17,217],"deployment.":[18],"In":[19],"this":[20],"work,":[21],"we":[22,134,173],"introduce":[23],"SDQ-LLM:":[24],"Sigma-Delta":[25,91],"Quantization":[26],"1-bit":[28,103],"LLMs":[29,42],"of":[30,41,52,58,131,147,157],"any":[31],"size,":[32],"a":[33,175,215],"novel":[34],"framework":[35],"that":[36,212],"enables":[37],"while":[43],"preserving":[44],"linguistic":[46],"reasoning":[47],"capabilities.":[48],"A":[49],"distinctive":[50],"feature":[51],"SDQ-LLM":[53,86,213],"is":[54,229],"the":[55,59,108,129,143,148,155,158,165],"continuous":[56],"adjustability":[57],"Over-Sampling":[60],"Ratio":[61],"(OSR),":[62],"enabling":[63],"dynamic":[64],"adaptation":[65],"to":[66,93,140,152],"or":[68,95,104],"VRAM":[69],"constraints":[70],"by":[71],"selecting":[72],"fractional":[73],"OSR":[74,159,180,187],"(e.g.,":[75],"2.5\u00d7)":[76],"an":[78],"optimal":[79],"trade-off":[80],"between":[81,167],"model":[82,209],"size":[83],"accuracy.":[85],"uses":[87],"upsampling":[88],"combined":[89],"with":[90,114],"Quantizer":[92],"binarize":[94],"ternarize":[96],"LLMs\u2019":[97],"weights,":[98],"encoding":[99],"high-precision":[100,219],"parameters":[101],"into":[102],"1.58-bit":[105],"representations,":[106],"replacing":[107],"multiplication":[109],"operations":[110],"within":[111,192],"linear":[112],"layers":[113,190],"addition.":[115],"This":[116,184],"approach":[117],"significantly":[118],"enhances":[119],"inference":[120],"efficiency":[121],"under":[122,222],"quantization.":[125],"To":[126],"further":[127],"reduce":[128,161],"loss":[130],"precision,":[133],"incorporate":[135],"Hadamard-based":[136],"weight":[137,149,171,197],"smoothing":[138],"prior":[139],"quantization,":[141],"improving":[142],"stability":[144],"robustness":[146],"representations.":[150],"Furthermore,":[151],"fully":[153],"leverage":[154],"continuity":[156],"precision":[162],"loss,":[163],"recognizing":[164],"correlation":[166],"sensitivity":[169],"variance,":[172],"propose":[174],"fine-grained,":[176],"layer-":[177],"linear-wise":[179],"allocation":[181],"strategy,":[182],"MultiOSR.":[183],"strategy":[185],"distributes":[186],"both":[188],"across":[189],"each":[193],"layer,":[194],"based":[195],"on":[196,205],"variance":[198],"parameter":[200],"scale.":[201],"Finally,":[202],"extensive":[203],"experiments":[204],"OPT":[206],"LLaMA":[208],"families":[210],"demonstrate":[211],"achieves":[214],"more":[216],"performance":[220],"even":[221],"highly":[223],"aggressive":[224],"low-OSR":[225],"settings.":[226],"Our":[227],"code":[228],"available":[230],"at":[231],"https://github.com/Dreamlittlecat/LLM-Quant-Factory.":[232]},"counts_by_year":[],"updated_date":"2026-03-07T16:01:11.037858","created_date":"2025-10-24T00:00:00"}