{"id":"https://openalex.org/W7166802776","doi":"https://doi.org/10.18653/v1/2026.findings-acl.1566","title":"pQuant: Towards Effective Low-Bit Language Models via Decoupled Linear Quantization-Aware Training","display_name":"pQuant: Towards Effective Low-Bit Language Models via Decoupled Linear Quantization-Aware Training","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7166802776","doi":"https://doi.org/10.18653/v1/2026.findings-acl.1566"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2026.findings-acl.1566","doi":"https://doi.org/10.18653/v1/2026.findings-acl.1566","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.1566","pdf_url":"https://aclanthology.org/2026.findings-acl.1566.pdf","source":{"id":"https://openalex.org/S4363605144","display_name":"Findings of the Association for Computational Linguistics: ACL (Findings of ACL)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: ACL 2026","raw_type":"proceedings-article"},"volume":null,"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"hybrid","oa_url":"https://aclanthology.org/2026.findings-acl.1566.pdf","any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5127019312","display_name":"Wenzheng Zhang","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":"Wenzheng Zhang","raw_affiliation_strings":["School of Computer Science , Peking University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Computer Science , Peking University","institution_ids":["https://openalex.org/I20231570"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5139751110","display_name":"Bingzheng Liu","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I24943067","display_name":"Fudan University","ror":"https://ror.org/013q1eq08","country_code":"CN","type":"education","lineage":["https://openalex.org/I24943067"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Bingzheng Liu","raw_affiliation_strings":["College of Future Information Technology , Fudan University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"College of Future Information Technology , Fudan University","institution_ids":["https://openalex.org/I24943067"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.47661306,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"31334","last_page":"31351"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.9750999808311462,"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":0.9750999808311462,"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":0.004100000020116568,"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/T10363","display_name":"Low-power high-performance VLSI design","score":0.002400000113993883,"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-aware-training","display_name":"quantization-aware training","score":0.949999988079071},{"id":"https://openalex.org/keywords/one-bit-quantization","display_name":"1-bit quantization","score":0.7049999833106995},{"id":"https://openalex.org/keywords/feature-scaling","display_name":"feature scaling","score":0.6259999871253967},{"id":"https://openalex.org/keywords/capacity-scaling","display_name":"capacity scaling","score":0.8460000157356262},{"id":"https://openalex.org/keywords/edge-deployment","display_name":"edge deployment","score":0.722000002861023},{"id":"https://openalex.org/keywords/large-language-models","display_name":"large language models","score":0.8069999814033508}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6093000173568726},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.5425000190734863},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4253000020980835},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.3366999924182892},{"id":"https://openalex.org/C163175372","wikidata":"https://www.wikidata.org/wiki/Q3339222","display_name":"Linear model","level":2,"score":0.29510000348091125},{"id":"https://openalex.org/C112972136","wikidata":"https://www.wikidata.org/wiki/Q7595718","display_name":"Stability (learning theory)","level":2,"score":0.2856000065803528},{"id":"https://openalex.org/C116834253","wikidata":"https://www.wikidata.org/wiki/Q2039217","display_name":"Identification (biology)","level":2,"score":0.2768999934196472},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.2556000053882599},{"id":"https://openalex.org/C177148314","wikidata":"https://www.wikidata.org/wiki/Q170084","display_name":"Generalization","level":2,"score":0.24199999868869781},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.2402999997138977}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2026.findings-acl.1566","doi":"https://doi.org/10.18653/v1/2026.findings-acl.1566","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.1566","pdf_url":"https://aclanthology.org/2026.findings-acl.1566.pdf","source":{"id":"https://openalex.org/S4363605144","display_name":"Findings of the Association for Computational Linguistics: ACL (Findings of ACL)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: ACL 2026","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2026.findings-acl.1566","doi":"https://doi.org/10.18653/v1/2026.findings-acl.1566","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.1566","pdf_url":"https://aclanthology.org/2026.findings-acl.1566.pdf","source":{"id":"https://openalex.org/S4363605144","display_name":"Findings of the Association for Computational Linguistics: ACL (Findings of ACL)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: ACL 2026","raw_type":"proceedings-article"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7166802776.pdf","grobid_xml":"https://content.openalex.org/works/W7166802776.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4399115018","https://openalex.org/W7166810475","https://openalex.org/W7163610126","https://openalex.org/W4404783494","https://openalex.org/W4407764019","https://openalex.org/W4387635190","https://openalex.org/W7164150941","https://openalex.org/W7166868033","https://openalex.org/W4415797319","https://openalex.org/W4401661003"],"abstract_inverted_index":{"Quantization-Aware":[0],"Training":[1],"from":[2],"scratch":[3],"has":[4],"emerged":[5],"as":[6,49],"a":[7,45,50,68,81,89],"promising":[8],"approach":[9],"for":[10,28,85],"building":[11],"efficient":[12,86,125],"large":[13],"language":[14],"models":[15],"(LLMs)":[16],"with":[17],"extremely":[18,136],"low-bit":[19,137],"weights":[20],"(sub":[21],"2-bit),":[22],"which":[23],"can":[24],"offer":[25],"substantial":[26],"advantages":[27],"edge":[29],"deployment.However,":[30],"existing":[31],"methods":[32],"still":[33],"fail":[34],"to":[35,94,108,112],"achieve":[36],"satisfactory":[37],"accuracy":[38],"and":[39,88],"scalability.In":[40],"this":[41,118],"work,":[42],"we":[43,65,103,116],"identify":[44],"parameter":[46],"democratization":[47],"effect":[48],"key":[51],"bottleneck:":[52],"the":[53,96,106,113],"sensitivity":[54],"of":[55],"all":[56],"parameters":[57,72,111],"becomes":[58],"homogenized,":[59],"severely":[60],"limiting":[61],"expressivity.To":[62],"address":[63],"this,":[64],"propose":[66],"pQuant,":[67],"method":[69],"that":[70],"decouples":[71],"by":[73],"splitting":[74],"linear":[75],"layers":[76],"into":[77,120],"two":[78],"specialized":[79],"branches:":[80],"dominant":[82],"1-bit":[83],"branch":[84,92,119],"computation":[87],"compact":[90],"highprecision":[91,114],"dedicated":[93],"preserving":[95],"most":[97],"sensitive":[98,110],"parameters.Through":[99],"tailored":[100],"feature":[101],"scaling,":[102],"explicitly":[104],"guide":[105],"model":[107],"allocate":[109],"branch.Furthermore,":[115],"extend":[117],"multiple,":[121],"sparsely-activated":[122],"experts,":[123],"enabling":[124],"capacity":[126],"scaling.Extensive":[127],"experiments":[128],"indicate":[129],"our":[130],"pQuant":[131],"achieves":[132],"stateof-the-art":[133],"performance":[134],"in":[135],"quantization.":[138]},"counts_by_year":[],"updated_date":"2026-10-28T10:33:24.094632","created_date":"2026-07-02T00:00:00"}