{"id":"https://openalex.org/W7214701152","doi":"https://doi.org/10.48550/arxiv.2609.34259","title":"QuantaSpike: Short-Window Spike-Driven Quantization for Large Language Models","display_name":"QuantaSpike: Short-Window Spike-Driven Quantization for Large Language Models","publication_year":2026,"publication_date":"2026-09-28","ids":{"openalex":"https://openalex.org/W7214701152","doi":"https://doi.org/10.48550/arxiv.2609.34259"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2609.34259","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.34259","pdf_url":"https://arxiv.org/pdf/2609.34259","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","is_traditional":true,"indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2609.34259","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5101307404","display_name":"Bang Hu","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":"Hu, Bang","raw_affiliation_strings":["School of Computer Science , Fudan University , Shanghai , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Computer Science , Fudan University , Shanghai , China","institution_ids":["https://openalex.org/I24943067"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5153518008","display_name":"Guowei Zhu","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":"Zhu, Guowei","raw_affiliation_strings":["School of Computer Science , Fudan University , Shanghai , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Computer Science , Fudan University , Shanghai , China","institution_ids":["https://openalex.org/I24943067"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5113112983","display_name":"Changze Lv","orcid":"https://orcid.org/0009-0007-7710-4982","observed_orcids":["https://orcid.org/0009-0007-7710-4982"]},"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":"Lv, Changze","raw_affiliation_strings":["School of Computer Science , Fudan University , Shanghai , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Computer Science , Fudan University , Shanghai , China","institution_ids":["https://openalex.org/I24943067"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5153440243","display_name":"Xiaoqing Zheng","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":"Zheng, Xiaoqing","raw_affiliation_strings":["School of Computer Science , Fudan University , Shanghai , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Computer Science , Fudan University , Shanghai , China","institution_ids":["https://openalex.org/I24943067"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5058922709","display_name":"Fengzhe Zhang","orcid":"https://orcid.org/0000-0002-7584-1045","observed_orcids":["https://orcid.org/0000-0002-7584-1045"]},"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":"Zhang, Fengzhe","raw_affiliation_strings":["School of Computer Science , Fudan University , Shanghai , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Computer Science , Fudan University , Shanghai , China","institution_ids":["https://openalex.org/I24943067"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5153523546","display_name":"Fan Zhang","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":"Zhang, Fan","raw_affiliation_strings":["School of Computer Science , Fudan University , Shanghai , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Computer Science , Fudan University , Shanghai , China","institution_ids":["https://openalex.org/I24943067"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5153432474","display_name":"Wei Cao","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":"Cao, Wei","raw_affiliation_strings":["School of Computer Science , Fudan University , Shanghai , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Computer Science , Fudan University , Shanghai , China","institution_ids":["https://openalex.org/I24943067"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I24943067"],"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/T10502","display_name":"Advanced Memory and Neural Computing","score":0.9789999723434448,"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"}},"topics":[{"id":"https://openalex.org/T10502","display_name":"Advanced Memory and Neural Computing","score":0.9789999723434448,"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"}},{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.005400000140070915,"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/T10363","display_name":"Low-power high-performance VLSI design","score":0.001500000013038516,"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/spiking-neural-networks","display_name":"spiking neural networks","score":0.8489999771118164},{"id":"https://openalex.org/keywords/large-language-models","display_name":"large language models","score":0.9049999713897705},{"id":"https://openalex.org/keywords/energy-efficient-inference","display_name":"energy-efficient inference","score":0.7160000205039978},{"id":"https://openalex.org/keywords/multiply-accumulate","display_name":"multiply-accumulate operations","score":0.5920000076293945},{"id":"https://openalex.org/keywords/llm-inference-acceleration","display_name":"LLM inference acceleration","score":0.3449999988079071}],"concepts":[{"id":"https://openalex.org/C100279451","wikidata":"https://www.wikidata.org/wiki/Q372193","display_name":"Perplexity","level":3,"score":0.6888999938964844},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6801999807357788},{"id":"https://openalex.org/C28855332","wikidata":"https://www.wikidata.org/wiki/Q198099","display_name":"Quantization (signal processing)","level":2,"score":0.6105999946594238},{"id":"https://openalex.org/C155512373","wikidata":"https://www.wikidata.org/wiki/Q287450","display_name":"Residual","level":2,"score":0.5831000208854675},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.525600016117096},{"id":"https://openalex.org/C79337645","wikidata":"https://www.wikidata.org/wiki/Q779824","display_name":"Outlier","level":2,"score":0.4943999946117401},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.49410000443458557},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.41929998993873596},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.41290000081062317},{"id":"https://openalex.org/C39927690","wikidata":"https://www.wikidata.org/wiki/Q11197","display_name":"Logarithm","level":2,"score":0.40860000252723694},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.37439998984336853},{"id":"https://openalex.org/C204241405","wikidata":"https://www.wikidata.org/wiki/Q461499","display_name":"Transformation (genetics)","level":3,"score":0.3395000100135803},{"id":"https://openalex.org/C2984842247","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep neural networks","level":3,"score":0.31049999594688416},{"id":"https://openalex.org/C186370098","wikidata":"https://www.wikidata.org/wiki/Q442787","display_name":"Energy (signal processing)","level":2,"score":0.2962999939918518},{"id":"https://openalex.org/C2777735758","wikidata":"https://www.wikidata.org/wiki/Q817765","display_name":"Path (computing)","level":2,"score":0.2849999964237213},{"id":"https://openalex.org/C45493050","wikidata":"https://www.wikidata.org/wiki/Q7884934","display_name":"Unified Model","level":2,"score":0.2847000062465668},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.27799999713897705},{"id":"https://openalex.org/C111030470","wikidata":"https://www.wikidata.org/wiki/Q1430460","display_name":"Curse of dimensionality","level":2,"score":0.27410000562667847},{"id":"https://openalex.org/C2944601119","wikidata":"https://www.wikidata.org/wiki/Q43744058","display_name":"Residual neural network","level":3,"score":0.2718999981880188},{"id":"https://openalex.org/C195324797","wikidata":"https://www.wikidata.org/wiki/Q33742","display_name":"Natural language","level":2,"score":0.27140000462532043},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.27000001072883606},{"id":"https://openalex.org/C124584101","wikidata":"https://www.wikidata.org/wiki/Q1053266","display_name":"Multiplier (economics)","level":2,"score":0.2637999951839447},{"id":"https://openalex.org/C64452783","wikidata":"https://www.wikidata.org/wiki/Q1524945","display_name":"Ternary operation","level":2,"score":0.2606000006198883},{"id":"https://openalex.org/C205711294","wikidata":"https://www.wikidata.org/wiki/Q176953","display_name":"Rendering (computer graphics)","level":2,"score":0.2517000138759613}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2609.34259","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.34259","pdf_url":"https://arxiv.org/pdf/2609.34259","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.2609.34259","doi":"https://doi.org/10.48550/arxiv.2609.34259","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2609.34259","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2609.34259","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.34259","pdf_url":"https://arxiv.org/pdf/2609.34259","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/W7214701152.pdf","grobid_xml":"https://content.openalex.org/works/W7214701152.grobid-xml"},"referenced_works_count":9,"referenced_works":["https://openalex.org/W4224308101","https://openalex.org/W4229005866","https://openalex.org/W4322718191","https://openalex.org/W4362598949","https://openalex.org/W4384918448","https://openalex.org/W4401306886","https://openalex.org/W4407308597","https://openalex.org/W4414939355","https://openalex.org/W7128524751"],"related_works":["https://openalex.org/W4417298709","https://openalex.org/W4400517998","https://openalex.org/W7155035841","https://openalex.org/W7203686197","https://openalex.org/W4404916353","https://openalex.org/W7211895469","https://openalex.org/W7155218047","https://openalex.org/W7085100946","https://openalex.org/W7163129372","https://openalex.org/W4414430426"],"abstract_inverted_index":{"Large":[0],"language":[1],"models":[2,184,189],"(LLMs)":[3],"achieve":[4],"strong":[5],"performance":[6],"across":[7],"many":[8],"tasks":[9],"but":[10],"rely":[11],"on":[12,156,182,187],"dense":[13,148],"multiply-accumulate":[14],"(MAC)":[15],"operations":[16],"during":[17],"inference,":[18],"resulting":[19],"in":[20,32],"high":[21],"energy":[22,174],"cost.":[23],"Spiking":[24],"neural":[25],"networks":[26],"(SNNs)":[27],"offer":[28],"an":[29,194],"event-driven":[30],"alternative":[31],"which":[33],"synaptic":[34],"integration":[35],"uses":[36,75],"lightweight":[37],"accumulation.":[38],"However,":[39],"spike-driven":[40,62,139,198],"LLM":[41,140,201],"inference":[42],"remains":[43],"difficult":[44],"because":[45],"outlier-heavy":[46],"activations":[47],"typically":[48],"require":[49],"long":[50],"firing":[51,88,164],"windows":[52],"or":[53,132],"auxiliary":[54],"non-spiking":[55],"paths.":[56],"We":[57],"propose":[58],"QuantaSpike,":[59],"a":[60],"short-window":[61],"quantization":[63,141],"framework":[64],"for":[65,200],"LLMs":[66],"built":[67],"around":[68],"Logarithmic":[69],"Ternary":[70],"Integrate-and-Fire":[71],"(LTIF)":[72],"neurons.":[73],"LTIF":[74,109],"ternary":[76],"events":[77],"with":[78,98],"power-of-two":[79],"membrane-response":[80],"quanta,":[81],"improving":[82],"the":[83,121,153,161,173],"information":[84],"represented":[85],"by":[86,179],"each":[87],"step":[89],"while":[90],"retaining":[91],"shift-ACC-compatible":[92],"computation.":[93],"QuantaSpike":[94,129,171],"combines":[95],"this":[96],"neuron":[97],"group-adaptive":[99],"gain":[100],"and":[101,127,135,158,185,196],"selective":[102],"outlier":[103],"admission:":[104],"normal":[105],"values":[106],"use":[107],"residual":[108,123],"steps,":[110],"whereas":[111],"admitted":[112],"outliers":[113],"receive":[114],"one":[115,176],"additional":[116],"onset":[117],"spike":[118],"before":[119],"entering":[120],"same":[122,162],"dynamics.":[124],"Across":[125],"OPT":[126,183],"Llama-2,":[128],"achieves":[130],"state-of-the-art":[131],"competitive":[133],"perplexity":[134],"zero-shot":[136],"accuracy":[137],"among":[138],"methods.":[142],"It":[143],"also":[144],"transfers":[145],"to":[146,152,191],"newer":[147],"LLMs,":[149],"remaining":[150],"close":[151],"FP16":[154],"reference":[155],"Llama-3-8B":[157],"Qwen3-8B":[159],"under":[160],"four-step":[163],"window.":[165],"Analytical":[166],"linear-energy":[167],"projections":[168],"show":[169],"that":[170],"reduces":[172],"of":[175],"linear":[177],"transformation":[178],"about":[180],"$80.0\\%$":[181],"$67.1\\%$":[186],"Llama-2":[188],"relative":[190],"SpikeQuant,":[192],"providing":[193],"accurate":[195],"energy-efficient":[197],"path":[199],"inference.":[202]},"counts_by_year":[],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-09-30T00:00:00"}