{"id":"https://openalex.org/W4415708384","doi":"https://doi.org/10.1109/icme59968.2025.11209915","title":"SQ-Delta: Ultra-High Delta Compression for LLMs via Joint Sparsification-Quantization","display_name":"SQ-Delta: Ultra-High Delta Compression for LLMs via Joint Sparsification-Quantization","publication_year":2025,"publication_date":"2025-06-30","ids":{"openalex":"https://openalex.org/W4415708384","doi":"https://doi.org/10.1109/icme59968.2025.11209915"},"language":null,"primary_location":{"id":"doi:10.1109/icme59968.2025.11209915","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icme59968.2025.11209915","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 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/A5019028268","display_name":"Yanfeng Jiang","orcid":"https://orcid.org/0000-0002-5783-4175"},"institutions":[{"id":"https://openalex.org/I205237279","display_name":"Nankai University","ror":"https://ror.org/01y1kjr75","country_code":"CN","type":"education","lineage":["https://openalex.org/I205237279"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yanfeng Jiang","raw_affiliation_strings":["Nankai University,College of Computer Science,Tianjin,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Nankai University,College of Computer Science,Tianjin,China","institution_ids":["https://openalex.org/I205237279"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Zelan Yang","orcid":null},"institutions":[{"id":"https://openalex.org/I45928872","display_name":"Alibaba Group (China)","ror":"https://ror.org/00k642b80","country_code":"CN","type":"company","lineage":["https://openalex.org/I45928872"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zelan Yang","raw_affiliation_strings":["Alibaba Inc,Hangzhou,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Alibaba Inc,Hangzhou,China","institution_ids":["https://openalex.org/I45928872"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5111133708","display_name":"Bohua Chen","orcid":"https://orcid.org/0009-0004-8927-8902"},"institutions":[{"id":"https://openalex.org/I45928872","display_name":"Alibaba Group (China)","ror":"https://ror.org/00k642b80","country_code":"CN","type":"company","lineage":["https://openalex.org/I45928872"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Bohua Chen","raw_affiliation_strings":["Alibaba Inc,Hangzhou,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Alibaba Inc,Hangzhou,China","institution_ids":["https://openalex.org/I45928872"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5014511114","display_name":"Li Shen","orcid":null},"institutions":[{"id":"https://openalex.org/I45928872","display_name":"Alibaba Group (China)","ror":"https://ror.org/00k642b80","country_code":"CN","type":"company","lineage":["https://openalex.org/I45928872"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Shen Li","raw_affiliation_strings":["Alibaba Inc,Hangzhou,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Alibaba Inc,Hangzhou,China","institution_ids":["https://openalex.org/I45928872"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5018180824","display_name":"Yong Li","orcid":"https://orcid.org/0000-0001-8049-9128"},"institutions":[{"id":"https://openalex.org/I45928872","display_name":"Alibaba Group (China)","ror":"https://ror.org/00k642b80","country_code":"CN","type":"company","lineage":["https://openalex.org/I45928872"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yong Li","raw_affiliation_strings":["Alibaba Inc,Hangzhou,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Alibaba Inc,Hangzhou,China","institution_ids":["https://openalex.org/I45928872"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5100455375","display_name":"Tao Li","orcid":"https://orcid.org/0000-0003-1546-7888"},"institutions":[{"id":"https://openalex.org/I205237279","display_name":"Nankai University","ror":"https://ror.org/01y1kjr75","country_code":"CN","type":"education","lineage":["https://openalex.org/I205237279"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Tao Li","raw_affiliation_strings":["Nankai University,College of Computer Science,Tianjin,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Nankai University,College of Computer Science,Tianjin,China","institution_ids":["https://openalex.org/I205237279"]}]}],"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":false,"cited_by_count":0,"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/T10901","display_name":"Advanced Data Compression Techniques","score":0.657800018787384,"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/T10901","display_name":"Advanced Data Compression Techniques","score":0.657800018787384,"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/T11269","display_name":"Algorithms and Data Compression","score":0.16840000450611115,"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/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.04910000041127205,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"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/compression","display_name":"Compression (physics)","score":0.7894999980926514},{"id":"https://openalex.org/keywords/dropout","display_name":"Dropout (neural networks)","score":0.5080000162124634},{"id":"https://openalex.org/keywords/joint","display_name":"Joint (building)","score":0.49970000982284546},{"id":"https://openalex.org/keywords/code","display_name":"Code (set theory)","score":0.4544999897480011},{"id":"https://openalex.org/keywords/data-compression","display_name":"Data compression","score":0.4519999921321869},{"id":"https://openalex.org/keywords/delta","display_name":"Delta","score":0.43149998784065247},{"id":"https://openalex.org/keywords/lossless-compression","display_name":"Lossless compression","score":0.42989999055862427},{"id":"https://openalex.org/keywords/range","display_name":"Range (aeronautics)","score":0.42820000648498535}],"concepts":[{"id":"https://openalex.org/C180016635","wikidata":"https://www.wikidata.org/wiki/Q2712821","display_name":"Compression (physics)","level":2,"score":0.7894999980926514},{"id":"https://openalex.org/C2776145597","wikidata":"https://www.wikidata.org/wiki/Q25339462","display_name":"Dropout (neural networks)","level":2,"score":0.5080000162124634},{"id":"https://openalex.org/C18555067","wikidata":"https://www.wikidata.org/wiki/Q8375051","display_name":"Joint (building)","level":2,"score":0.49970000982284546},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.4578999876976013},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.4544999897480011},{"id":"https://openalex.org/C78548338","wikidata":"https://www.wikidata.org/wiki/Q2493","display_name":"Data compression","level":2,"score":0.4519999921321869},{"id":"https://openalex.org/C5072461","wikidata":"https://www.wikidata.org/wiki/Q49506","display_name":"Delta","level":2,"score":0.43149998784065247},{"id":"https://openalex.org/C81081738","wikidata":"https://www.wikidata.org/wiki/Q55542","display_name":"Lossless compression","level":3,"score":0.42989999055862427},{"id":"https://openalex.org/C204323151","wikidata":"https://www.wikidata.org/wiki/Q905424","display_name":"Range (aeronautics)","level":2,"score":0.42820000648498535},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.40560001134872437},{"id":"https://openalex.org/C25797200","wikidata":"https://www.wikidata.org/wiki/Q828137","display_name":"Compression ratio","level":3,"score":0.3930000066757202},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.3625999987125397},{"id":"https://openalex.org/C196083921","wikidata":"https://www.wikidata.org/wiki/Q7915758","display_name":"Variance (accounting)","level":2,"score":0.35499998927116394},{"id":"https://openalex.org/C94835093","wikidata":"https://www.wikidata.org/wiki/Q3113333","display_name":"Data compression ratio","level":5,"score":0.3359000086784363},{"id":"https://openalex.org/C25432639","wikidata":"https://www.wikidata.org/wiki/Q3711790","display_name":"Weight distribution","level":2,"score":0.3102000057697296},{"id":"https://openalex.org/C13481523","wikidata":"https://www.wikidata.org/wiki/Q412438","display_name":"Image compression","level":4,"score":0.2989000082015991},{"id":"https://openalex.org/C105795698","wikidata":"https://www.wikidata.org/wiki/Q12483","display_name":"Statistics","level":1,"score":0.2930000126361847},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.28760001063346863},{"id":"https://openalex.org/C66938386","wikidata":"https://www.wikidata.org/wiki/Q633538","display_name":"Structural engineering","level":1,"score":0.2874000072479248},{"id":"https://openalex.org/C2778156585","wikidata":"https://www.wikidata.org/wiki/Q174053","display_name":"Relay","level":3,"score":0.2524999976158142},{"id":"https://openalex.org/C105339364","wikidata":"https://www.wikidata.org/wiki/Q2297740","display_name":"Software deployment","level":2,"score":0.2515999972820282},{"id":"https://openalex.org/C148043351","wikidata":"https://www.wikidata.org/wiki/Q4456944","display_name":"Current (fluid)","level":2,"score":0.2502000033855438}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/icme59968.2025.11209915","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icme59968.2025.11209915","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE International Conference on Multimedia and Expo (ICME)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"}],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":5,"referenced_works":["https://openalex.org/W2979826702","https://openalex.org/W4385573610","https://openalex.org/W4386160289","https://openalex.org/W4387321091","https://openalex.org/W4412887834"],"related_works":[],"abstract_inverted_index":{"Current":[0],"delta":[1,31,54,76,89],"compression":[2,32,42,107,133,140],"methods":[3],"struggle":[4],"to":[5,10,39,71,95,112],"achieve":[6,96],"ultra-high":[7,129],"compression,":[8,130],"failing":[9],"minimize":[11],"the":[12,41,48,53,75,87,126,135,142],"deployment":[13],"costs":[14],"of":[15],"multiple":[16],"full-parameter":[17],"fine-tuned":[18],"large":[19],"language":[20],"models.":[21],"To":[22],"address":[23],"this":[24,64],"issue,":[25],"we":[26,67],"propose":[27],"a":[28,97],"novel":[29],"distribution-driven":[30],"framework":[33],"SQ-Delta,":[34],"which":[35],"utilizes":[36],"joint":[37],"sparsification-quantization":[38],"maximize":[40],"ratio.":[43],"We":[44],"have":[45],"observed":[46],"that":[47,103],"matrix-computed":[49],"intermediate":[50],"results":[51,101],"for":[52,114,128,134,141],"weight":[55,77,90],"exhibit":[56],"small":[57],"variance":[58],"and":[59,93,116,138],"min-max":[60],"range":[61],"characteristics.":[62],"Exploiting":[63],"distribution":[65],"phenomenon,":[66],"introduce":[68],"Group-wise":[69],"Dropout":[70],"perform":[72],"dropout":[73],"on":[74],"using":[78,84],"an":[79],"optimal":[80],"group":[81],"size.":[82],"Furthermore,":[83],"Separate":[85],"Quantization,":[86],"sparse":[88],"is":[91,147],"quantized":[92],"decomposed":[94],"lower":[98],"bit.":[99],"Experimental":[100],"show":[102],"SQ-Delta":[104,124],"achieves":[105],"16\u00d7":[106],"with":[108],"improved":[109],"accuracy":[110],"compared":[111],"baselines":[113],"WizardMath":[115],"WizardCoder":[117],"models":[118],"across":[119],"different":[120],"parameter":[121],"scales.":[122],"Moreover,":[123],"demonstrates":[125],"ability":[127],"achieving":[131],"128\u00d7":[132],"WizardMath-7B":[136],"model":[137],"512\u00d7":[139],"WizardMath-70B":[143],"model.":[144],"Our":[145],"code":[146],"available":[148],"at":[149],"https://github.com/llwx593/SQ-Delta.":[150]},"counts_by_year":[],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2025-10-30T00:00:00"}