{"id":"https://openalex.org/W4416251074","doi":"https://doi.org/10.1109/ijcnn64981.2025.11229416","title":"Low-Rank Decomposition Assisted Quantization and Inference Compensation for Quality Large Language Model Inference","display_name":"Low-Rank Decomposition Assisted Quantization and Inference Compensation for Quality Large Language Model Inference","publication_year":2025,"publication_date":"2025-06-30","ids":{"openalex":"https://openalex.org/W4416251074","doi":"https://doi.org/10.1109/ijcnn64981.2025.11229416"},"language":null,"primary_location":{"id":"doi:10.1109/ijcnn64981.2025.11229416","is_oa":false,"landing_page_url":"https://doi.org/10.1109/ijcnn64981.2025.11229416","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 International Joint Conference on Neural Networks (IJCNN)","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/A5057428751","display_name":"Jie Ou","orcid":"https://orcid.org/0000-0002-1159-2043"},"institutions":[{"id":"https://openalex.org/I150229711","display_name":"University of Electronic Science and Technology of China","ror":"https://ror.org/04qr3zq92","country_code":"CN","type":"education","lineage":["https://openalex.org/I150229711"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jie Ou","raw_affiliation_strings":["University of Electronic Science and Technology of China,School of Information and Software Engineering,Chengdu,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Electronic Science and Technology of China,School of Information and Software Engineering,Chengdu,China","institution_ids":["https://openalex.org/I150229711"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5103538643","display_name":"Jinyu Guo","orcid":null},"institutions":[{"id":"https://openalex.org/I150229711","display_name":"University of Electronic Science and Technology of China","ror":"https://ror.org/04qr3zq92","country_code":"CN","type":"education","lineage":["https://openalex.org/I150229711"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jinyu Guo","raw_affiliation_strings":["University of Electronic Science and Technology of China,School of Information and Software Engineering,Chengdu,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Electronic Science and Technology of China,School of Information and Software Engineering,Chengdu,China","institution_ids":["https://openalex.org/I150229711"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5111430408","display_name":"Shuaihong Jiang","orcid":null},"institutions":[{"id":"https://openalex.org/I150229711","display_name":"University of Electronic Science and Technology of China","ror":"https://ror.org/04qr3zq92","country_code":"CN","type":"education","lineage":["https://openalex.org/I150229711"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Shuaihong Jiang","raw_affiliation_strings":["University of Electronic Science and Technology of China,School of Information and Software Engineering,Chengdu,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Electronic Science and Technology of China,School of Information and Software Engineering,Chengdu,China","institution_ids":["https://openalex.org/I150229711"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5033745915","display_name":"Zhaokun Wang","orcid":"https://orcid.org/0000-0003-1937-7046"},"institutions":[{"id":"https://openalex.org/I150229711","display_name":"University of Electronic Science and Technology of China","ror":"https://ror.org/04qr3zq92","country_code":"CN","type":"education","lineage":["https://openalex.org/I150229711"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhaokun Wang","raw_affiliation_strings":["University of Electronic Science and Technology of China,School of Information and Software Engineering,Chengdu,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Electronic Science and Technology of China,School of Information and Software Engineering,Chengdu,China","institution_ids":["https://openalex.org/I150229711"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5033702874","display_name":"Yueming Chen","orcid":null},"institutions":[{"id":"https://openalex.org/I150229711","display_name":"University of Electronic Science and Technology of China","ror":"https://ror.org/04qr3zq92","country_code":"CN","type":"education","lineage":["https://openalex.org/I150229711"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yueming Chen","raw_affiliation_strings":["University of Electronic Science and Technology of China,School of Information and Software Engineering,Chengdu,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Electronic Science and Technology of China,School of Information and Software Engineering,Chengdu,China","institution_ids":["https://openalex.org/I150229711"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5047127750","display_name":"Ruini Xue","orcid":"https://orcid.org/0000-0003-1802-5188"},"institutions":[{"id":"https://openalex.org/I150229711","display_name":"University of Electronic Science and Technology of China","ror":"https://ror.org/04qr3zq92","country_code":"CN","type":"education","lineage":["https://openalex.org/I150229711"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Ruini Xue","raw_affiliation_strings":["University of Electronic Science and Technology of China,School of Computer Science and Engineering,Chengdu,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Electronic Science and Technology of China,School of Computer Science and Engineering,Chengdu,China","institution_ids":["https://openalex.org/I150229711"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5001104882","display_name":"Wenhong Tian","orcid":"https://orcid.org/0000-0002-5551-9796"},"institutions":[{"id":"https://openalex.org/I150229711","display_name":"University of Electronic Science and Technology of China","ror":"https://ror.org/04qr3zq92","country_code":"CN","type":"education","lineage":["https://openalex.org/I150229711"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Wenhong Tian","raw_affiliation_strings":["University of Electronic Science and Technology of China,School of Information and Software Engineering,Chengdu,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Electronic Science and Technology of China,School of Information and Software Engineering,Chengdu,China","institution_ids":["https://openalex.org/I150229711"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I150229711"],"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":"9"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.2718999981880188,"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":0.2718999981880188,"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":0.08940000087022781,"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":0.08569999784231186,"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/inference","display_name":"Inference","score":0.8607000112533569},{"id":"https://openalex.org/keywords/quantization","display_name":"Quantization (signal processing)","score":0.7644000053405762},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.4510999917984009},{"id":"https://openalex.org/keywords/compensation","display_name":"Compensation (psychology)","score":0.38989999890327454},{"id":"https://openalex.org/keywords/quality","display_name":"Quality (philosophy)","score":0.30140000581741333}],"concepts":[{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.8607000112533569},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.76910001039505},{"id":"https://openalex.org/C28855332","wikidata":"https://www.wikidata.org/wiki/Q198099","display_name":"Quantization (signal processing)","level":2,"score":0.7644000053405762},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4968999922275543},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.4510999917984009},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4146000146865845},{"id":"https://openalex.org/C2780023022","wikidata":"https://www.wikidata.org/wiki/Q1338171","display_name":"Compensation (psychology)","level":2,"score":0.38989999890327454},{"id":"https://openalex.org/C2779530757","wikidata":"https://www.wikidata.org/wiki/Q1207505","display_name":"Quality (philosophy)","level":2,"score":0.30140000581741333},{"id":"https://openalex.org/C195324797","wikidata":"https://www.wikidata.org/wiki/Q33742","display_name":"Natural language","level":2,"score":0.30000001192092896},{"id":"https://openalex.org/C124681953","wikidata":"https://www.wikidata.org/wiki/Q339062","display_name":"Decomposition","level":2,"score":0.29989999532699585},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.27709999680519104},{"id":"https://openalex.org/C186108316","wikidata":"https://www.wikidata.org/wiki/Q352530","display_name":"Adaptive neuro fuzzy inference system","level":4,"score":0.2720000147819519}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/ijcnn64981.2025.11229416","is_oa":false,"landing_page_url":"https://doi.org/10.1109/ijcnn64981.2025.11229416","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 International Joint Conference on Neural Networks (IJCNN)","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"},{"id":"https://openalex.org/F4320326670","display_name":"Chengdu Science and Technology Bureau","ror":null}],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":11,"referenced_works":["https://openalex.org/W2093647425","https://openalex.org/W2946609015","https://openalex.org/W2998617917","https://openalex.org/W3194676777","https://openalex.org/W4321636575","https://openalex.org/W4367672983","https://openalex.org/W4389524393","https://openalex.org/W4393145687","https://openalex.org/W4401043095","https://openalex.org/W4402670692","https://openalex.org/W4406650295"],"related_works":[],"abstract_inverted_index":{"Large":[0],"Language":[1],"Models":[2],"(LLMs)":[3],"have":[4],"demonstrated":[5],"exceptional":[6],"performance":[7,69],"on":[8,102,162],"natural":[9],"language":[10],"processing":[11],"tasks.":[12],"However,":[13],"these":[14],"models":[15],"are":[16],"computationally":[17],"intensive":[18],"and":[19,118,181],"require":[20],"substantial":[21],"hardware":[22],"resources":[23],"for":[24,33,104],"deployment.":[25],"Quantization":[26,78],"has":[27],"emerged":[28],"as":[29],"a":[30,60],"popular":[31],"technique":[32],"LLM":[34],"deployment,":[35],"reducing":[36],"memory":[37],"requirements,":[38],"but":[39],"it":[40],"results":[41,151],"in":[42],"accuracy":[43,53,105,145],"degradation,":[44],"particularly":[45],"when":[46],"using":[47],"low-bit":[48],"quantization.":[49,73],"To":[50],"mitigate":[51],"this":[52],"loss,":[54],"we":[55,75,128],"introduce":[56],"Low-Rank":[57,77],"Compensation":[58],"(LoRC),":[59],"novel":[61],"compensation":[62,112],"mechanism":[63],"that":[64,153],"aims":[65],"to":[66,94,135],"recover":[67],"the":[68,83,91,115,130,139,157,168,172,176,190],"drop":[70],"caused":[71],"by":[72,86,160],"Additionally,":[74],"propose":[76,129],"(LoRQ),":[79],"which":[80,142],"further":[81],"reduces":[82],"quantization-induced":[84],"loss":[85,106],"adaptively":[87],"adjusting":[88],"weights":[89],"at":[90],"element-wise":[92],"level":[93],"help":[95],"LLMs":[96],"accommodate":[97],"quantized":[98],"computations.":[99],"LoRC":[100,137],"focuses":[101],"compensating":[103],"during":[107],"inference,":[108],"LoRQ":[109],"integrates":[110],"low-rank":[111],"directly":[113],"into":[114,138,189],"quantization":[116],"process,":[117],"they":[119],"do":[120],"not":[121],"need":[122],"end-to-end":[123],"fine-tuning":[124],"with":[125],"LLM.":[126],"Furthermore,":[127],"Rank-\u03b1":[131],"Addition":[132],"Strategy":[133],"(R\u03b1AS)":[134],"combine":[136],"inference":[140,144,148,192,197],"framework,":[141],"improves":[143],"without":[146,194],"increasing":[147],"latency.":[149],"Experimental":[150],"show":[152],"our":[154,183],"method":[155],"outperforms":[156],"state-of-the-art":[158],"OmniQuant":[159],"1.89%":[161],"several":[163],"common":[164],"zero-shot":[165],"datasets":[166],"under":[167],"W4A4":[169],"setting":[170],"of":[171,179],"widely-used":[173],"LLaMA.":[174],"Through":[175],"joint":[177],"design":[178],"algorithms":[180],"systems,":[182],"techniques":[184],"can":[185],"be":[186],"easily":[187],"integrated":[188],"FlexGen":[191],"framework":[193],"introducing":[195],"additional":[196],"latency,":[198],"thereby":[199],"maintaining":[200],"high":[201],"throughput":[202],"while":[203],"improving":[204],"accuracy.":[205]},"counts_by_year":[],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2025-11-14T00:00:00"}