{"id":"https://openalex.org/W7224077285","doi":"https://doi.org/10.48550/arxiv.2610.37336","title":"CoS-VQ: Coordinated Shared-Scale Vector Quantization for Large Language Models","display_name":"CoS-VQ: Coordinated Shared-Scale Vector Quantization for Large Language Models","publication_year":2026,"publication_date":"2026-10-27","ids":{"openalex":"https://openalex.org/W7224077285","doi":"https://doi.org/10.48550/arxiv.2610.37336"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2610.37336","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.37336","pdf_url":"https://arxiv.org/pdf/2610.37336","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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"},"volume":null,"type":"preprint","indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2610.37336","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5160054672","display_name":"Zukang Xu","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xu, Zukang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5125940974","display_name":"Zhixiong Zhao","orcid":null,"observed_orcids":[]},"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":"Zhao, Zhixiong","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/A5160017594","display_name":"Xing Hu","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I76569877","display_name":"Southeast University","ror":"https://ror.org/04ct4d772","country_code":"CN","type":"education","lineage":["https://openalex.org/I76569877"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Hu, Xing","raw_affiliation_strings":["Southeast University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Southeast University","institution_ids":["https://openalex.org/I76569877"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5160048620","display_name":"Zhe Jiang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I76569877","display_name":"Southeast University","ror":"https://ror.org/04ct4d772","country_code":"CN","type":"education","lineage":["https://openalex.org/I76569877"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jiang, Zhe","raw_affiliation_strings":["Southeast University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Southeast University","institution_ids":["https://openalex.org/I76569877"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5159992704","display_name":"Xin Si","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Si, Xin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5160014651","display_name":"Dawei Yang","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yang, Dawei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"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":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.86080002784729,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.86080002784729,"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/T10028","display_name":"Topic Modeling","score":0.04399999976158142,"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/T10901","display_name":"Advanced Data Compression Techniques","score":0.023000000044703484,"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/vector-quantization","display_name":"vector quantization","score":0.9800000190734863},{"id":"https://openalex.org/keywords/large-language-models","display_name":"large language models","score":0.8220000267028809}],"concepts":[{"id":"https://openalex.org/C153207627","wikidata":"https://www.wikidata.org/wiki/Q863873","display_name":"Code word","level":3,"score":0.8233000040054321},{"id":"https://openalex.org/C199833920","wikidata":"https://www.wikidata.org/wiki/Q612536","display_name":"Vector quantization","level":2,"score":0.6484000086784363},{"id":"https://openalex.org/C28855332","wikidata":"https://www.wikidata.org/wiki/Q198099","display_name":"Quantization (signal processing)","level":2,"score":0.5228999853134155},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.5210000276565552},{"id":"https://openalex.org/C114466953","wikidata":"https://www.wikidata.org/wiki/Q6034165","display_name":"Initialization","level":2,"score":0.5144000053405762},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5047000050544739},{"id":"https://openalex.org/C129844170","wikidata":"https://www.wikidata.org/wiki/Q41299","display_name":"Quadratic equation","level":2,"score":0.4957999885082245},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.46779999136924744},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.44909998774528503},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.4397999942302704},{"id":"https://openalex.org/C179518139","wikidata":"https://www.wikidata.org/wiki/Q5140297","display_name":"Coding (social sciences)","level":2,"score":0.4262999892234802},{"id":"https://openalex.org/C2777984123","wikidata":"https://www.wikidata.org/wiki/Q9248237","display_name":"Pointwise","level":2,"score":0.4189999997615814},{"id":"https://openalex.org/C77553402","wikidata":"https://www.wikidata.org/wiki/Q13222579","display_name":"Upper and lower bounds","level":2,"score":0.38940000534057617},{"id":"https://openalex.org/C163716315","wikidata":"https://www.wikidata.org/wiki/Q901177","display_name":"Gaussian","level":2,"score":0.353300005197525},{"id":"https://openalex.org/C72169020","wikidata":"https://www.wikidata.org/wiki/Q194404","display_name":"Monotonic function","level":2,"score":0.33869999647140503},{"id":"https://openalex.org/C138187205","wikidata":"https://www.wikidata.org/wiki/Q131251","display_name":"Tangent","level":2,"score":0.3034999966621399},{"id":"https://openalex.org/C43126263","wikidata":"https://www.wikidata.org/wiki/Q128751","display_name":"Source code","level":2,"score":0.30300000309944153},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.2937999963760376},{"id":"https://openalex.org/C2778755073","wikidata":"https://www.wikidata.org/wiki/Q10858537","display_name":"Scale (ratio)","level":2,"score":0.28349998593330383},{"id":"https://openalex.org/C127759330","wikidata":"https://www.wikidata.org/wiki/Q637416","display_name":"Codebook","level":2,"score":0.28209999203681946},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.27549999952316284},{"id":"https://openalex.org/C68693459","wikidata":"https://www.wikidata.org/wiki/Q657586","display_name":"Hyperplane","level":2,"score":0.2745000123977661},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.2727000117301941},{"id":"https://openalex.org/C95925971","wikidata":"https://www.wikidata.org/wiki/Q3100414","display_name":"Parity-check matrix","level":4,"score":0.26350000500679016},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.26080000400543213},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.26030001044273376},{"id":"https://openalex.org/C70710897","wikidata":"https://www.wikidata.org/wiki/Q680081","display_name":"Separable space","level":2,"score":0.2567000091075897},{"id":"https://openalex.org/C2776036281","wikidata":"https://www.wikidata.org/wiki/Q48769818","display_name":"Constraint (computer-aided design)","level":2,"score":0.25540000200271606},{"id":"https://openalex.org/C118615104","wikidata":"https://www.wikidata.org/wiki/Q121416","display_name":"Discrete mathematics","level":1,"score":0.2533999979496002}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2610.37336","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.37336","pdf_url":"https://arxiv.org/pdf/2610.37336","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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"},{"id":"doi:10.48550/arxiv.2610.37336","doi":"https://doi.org/10.48550/arxiv.2610.37336","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2610.37336","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2610.37336","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.37336","pdf_url":"https://arxiv.org/pdf/2610.37336","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7224077285.pdf","grobid_xml":"https://content.openalex.org/works/W7224077285.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W7210287623","https://openalex.org/W4417095999","https://openalex.org/W4404782943","https://openalex.org/W4414820584","https://openalex.org/W7223335015","https://openalex.org/W4409248487","https://openalex.org/W4404261609","https://openalex.org/W7164197693","https://openalex.org/W4399317903","https://openalex.org/W4414580990"],"abstract_inverted_index":{"Large":[0],"language":[1,147],"models":[2],"(LLMs)":[3],"require":[4],"substantial":[5],"memory,":[6],"motivating":[7],"low-bit":[8],"weight":[9],"quantization":[10,16],"for":[11,105,128],"efficient":[12],"deployment.":[13],"Existing":[14],"vector":[15],"(VQ)":[17],"methods":[18],"often":[19],"improve":[20],"accuracy":[21],"by":[22],"increasing":[23,176],"the":[24,65,70,106,145,155],"effective":[25],"coding":[26],"dimension,":[27],"which":[28,44],"can":[29,85],"complicate":[30],"independent":[31],"lookup":[32],"and":[33,53,98,142,150],"local":[34],"codeword":[35,72,135,177],"updates.":[36],"We":[37],"propose":[38],"CoS-VQ":[39,91,139],"(Coordinated":[40],"Shared-Scale":[41],"Vector":[42],"Quantization),":[43],"preserves":[45],"low-dimensional":[46],"codewords":[47],"while":[48],"jointly":[49],"optimizing":[50],"hard":[51],"indices":[52],"their":[54],"shared":[55],"scale":[56,67,89],"across":[57,80],"positions.":[58],"The":[59,179],"key":[60],"challenge":[61],"is":[62,181],"scale-induced":[63],"coupling:":[64],"optimal":[66],"depends":[68],"on":[69,144],"entire":[71],"combination":[73],"and,":[74],"in":[75],"turn,":[76],"changes":[77],"candidate":[78,112],"rankings":[79],"positions,":[81],"so":[82],"pointwise":[83],"updates":[84,131,164],"stall":[86],"even":[87],"after":[88],"re-optimization.":[90],"collects":[92],"candidates":[93],"at":[94,183],"multiple":[95],"probe":[96],"scales":[97],"constructs":[99],"a":[100],"tangent":[101],"separable":[102],"upper":[103],"bound":[104],"residual-aware":[107],"quadratic":[108],"objective.":[109],"By":[110],"reducing":[111],"comparisons":[113],"to":[114,175],"lower":[115],"envelopes":[116],"of":[117],"lines":[118],"over":[119],"positive":[120],"scales,":[121],"it":[122],"identifies":[123],"finitely":[124],"many":[125],"representative":[126],"intervals":[127],"joint":[129,163,170],"index--scale":[130],"without":[132],"exhaustively":[133],"enumerating":[134],"combinations.":[136],"On":[137],"Qwen3.8-27B,":[138],"outperforms":[140],"QTIP":[141],"LLVQ":[143],"evaluated":[146],"modeling,":[148],"coding,":[149],"reasoning":[151],"benchmarks.":[152],"Ablations":[153],"from":[154,162,166],"same":[156],"initialization":[157],"further":[158],"show":[159],"larger":[160],"gains":[161],"than":[165],"scale-only":[167],"refinement,":[168],"highlighting":[169],"selection":[171],"as":[172],"an":[173],"alternative":[174],"dimensionality.":[178],"code":[180],"available":[182],"https://github.com/xzktx003/Cos_VQ.":[184]},"counts_by_year":[],"updated_date":"2026-10-31T12:09:11.096373","created_date":"2026-10-29T00:00:00"}