{"id":"https://openalex.org/W7162769050","doi":"https://doi.org/10.48550/arxiv.2605.29157","title":"Parallax: Parameterized Local Linear Attention for Language Modeling","display_name":"Parallax: Parameterized Local Linear Attention for Language Modeling","publication_year":2026,"publication_date":"2026-05-27","ids":{"openalex":"https://openalex.org/W7162769050","doi":"https://doi.org/10.48550/arxiv.2605.29157"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2605.29157","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.29157","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2605.29157","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5137362285","display_name":"Yifei Zuo","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zuo, Yifei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5120251973","display_name":"Dhruv Pai","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Pai, Dhruv","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137377000","display_name":"Zhichen Zeng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zeng, Zhichen","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137394019","display_name":"Alec Dewulf","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Dewulf, Alec","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137331277","display_name":"Shuming Hu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hu, Shuming","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5137338576","display_name":"Zhaoran Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Zhaoran","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"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":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11948","display_name":"Machine Learning in Materials Science","score":0.21649999916553497,"subfield":{"id":"https://openalex.org/subfields/2505","display_name":"Materials Chemistry"},"field":{"id":"https://openalex.org/fields/25","display_name":"Materials Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T11948","display_name":"Machine Learning in Materials Science","score":0.21649999916553497,"subfield":{"id":"https://openalex.org/subfields/2505","display_name":"Materials Chemistry"},"field":{"id":"https://openalex.org/fields/25","display_name":"Materials Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.14020000398159027,"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":0.10509999841451645,"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/kernelization","display_name":"Kernelization","score":0.46540001034736633},{"id":"https://openalex.org/keywords/context","display_name":"Context (archaeology)","score":0.45339998602867126},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.41290000081062317},{"id":"https://openalex.org/keywords/parameterized-complexity","display_name":"Parameterized complexity","score":0.39500001072883606},{"id":"https://openalex.org/keywords/convolutional-neural-network","display_name":"Convolutional neural network","score":0.385699987411499},{"id":"https://openalex.org/keywords/contrast","display_name":"Contrast (vision)","score":0.37880000472068787},{"id":"https://openalex.org/keywords/scalability","display_name":"Scalability","score":0.37599998712539673},{"id":"https://openalex.org/keywords/solver","display_name":"Solver","score":0.3709999918937683},{"id":"https://openalex.org/keywords/overfitting","display_name":"Overfitting","score":0.37059998512268066}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6601999998092651},{"id":"https://openalex.org/C207225210","wikidata":"https://www.wikidata.org/wiki/Q1759539","display_name":"Kernelization","level":3,"score":0.46540001034736633},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4648999869823456},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.45899999141693115},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.45339998602867126},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.4154999852180481},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.41290000081062317},{"id":"https://openalex.org/C165464430","wikidata":"https://www.wikidata.org/wiki/Q1570441","display_name":"Parameterized complexity","level":2,"score":0.39500001072883606},{"id":"https://openalex.org/C81363708","wikidata":"https://www.wikidata.org/wiki/Q17084460","display_name":"Convolutional neural network","level":2,"score":0.385699987411499},{"id":"https://openalex.org/C2776502983","wikidata":"https://www.wikidata.org/wiki/Q690182","display_name":"Contrast (vision)","level":2,"score":0.37880000472068787},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.37599998712539673},{"id":"https://openalex.org/C2778770139","wikidata":"https://www.wikidata.org/wiki/Q1966904","display_name":"Solver","level":2,"score":0.3709999918937683},{"id":"https://openalex.org/C22019652","wikidata":"https://www.wikidata.org/wiki/Q331309","display_name":"Overfitting","level":3,"score":0.37059998512268066},{"id":"https://openalex.org/C163175372","wikidata":"https://www.wikidata.org/wiki/Q3339222","display_name":"Linear model","level":2,"score":0.34040001034736633},{"id":"https://openalex.org/C74193536","wikidata":"https://www.wikidata.org/wiki/Q574844","display_name":"Kernel (algebra)","level":2,"score":0.33869999647140503},{"id":"https://openalex.org/C92757383","wikidata":"https://www.wikidata.org/wiki/Q382497","display_name":"Affine transformation","level":2,"score":0.3215999901294708},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.321399986743927},{"id":"https://openalex.org/C188441871","wikidata":"https://www.wikidata.org/wiki/Q7554146","display_name":"Softmax function","level":3,"score":0.31690001487731934},{"id":"https://openalex.org/C173801870","wikidata":"https://www.wikidata.org/wiki/Q201413","display_name":"Heuristic","level":2,"score":0.31369999051094055},{"id":"https://openalex.org/C2779231336","wikidata":"https://www.wikidata.org/wiki/Q7534724","display_name":"Sketch","level":2,"score":0.30480000376701355},{"id":"https://openalex.org/C159423971","wikidata":"https://www.wikidata.org/wiki/Q177251","display_name":"Associative property","level":2,"score":0.3037000000476837},{"id":"https://openalex.org/C112972136","wikidata":"https://www.wikidata.org/wiki/Q7595718","display_name":"Stability (learning theory)","level":2,"score":0.3001999855041504},{"id":"https://openalex.org/C127705205","wikidata":"https://www.wikidata.org/wiki/Q5748245","display_name":"Heuristics","level":2,"score":0.299699991941452},{"id":"https://openalex.org/C189645446","wikidata":"https://www.wikidata.org/wiki/Q350865","display_name":"Mirroring","level":2,"score":0.2775999903678894},{"id":"https://openalex.org/C102366305","wikidata":"https://www.wikidata.org/wiki/Q1097688","display_name":"Nonparametric statistics","level":2,"score":0.27630001306533813},{"id":"https://openalex.org/C15759828","wikidata":"https://www.wikidata.org/wiki/Q165074","display_name":"Parallax","level":2,"score":0.27489998936653137},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.2630000114440918},{"id":"https://openalex.org/C28855332","wikidata":"https://www.wikidata.org/wiki/Q198099","display_name":"Quantization (signal processing)","level":2,"score":0.25769999623298645},{"id":"https://openalex.org/C137800194","wikidata":"https://www.wikidata.org/wiki/Q11713455","display_name":"Interpolation (computer graphics)","level":3,"score":0.25619998574256897}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2605.29157","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.29157","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2605.29157","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.29157","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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"},"sustainable_development_goals":[{"display_name":"Quality Education","score":0.40747126936912537,"id":"https://metadata.un.org/sdg/4"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Large":[0],"Language":[1],"Models":[2],"(LLMs)":[3],"have":[4],"become":[5],"the":[6,13,36,51,102,115,129,131,135,145,222,231,243],"central":[7],"paradigm":[8],"in":[9,35,55,77,105,242],"artificial":[10],"intelligence,":[11],"yet":[12],"core":[14],"computational":[15,82],"primitive":[16],"of":[17,124,224,235],"attention":[18,29,47,57,125,151,240],"has":[19,73],"remained":[20],"structurally":[21],"unchanged.":[22],"Local":[23,92],"Linear":[24,93],"Attention":[25,94],"(LLA)":[26],"is":[27,96,230],"an":[28,109],"mechanism":[30],"derived":[31],"from":[32],"nonparametric":[33],"statistics":[34],"test-time":[37],"regression":[38],"framework.":[39],"In":[40],"contrast":[41],"to":[42,58,81,193],"prior":[43],"research":[44,245],"on":[45],"efficient":[46],"variants,":[48],"LLA":[49,72,106],"upgrades":[50],"local":[52,60],"constant":[53],"estimate":[54],"softmax":[56],"a":[59,90,122,140,153,206,216],"linear":[61],"estimate,":[62],"yielding":[63],"provably":[64],"superior":[65],"bias-variance":[66],"tradeoffs":[67],"for":[68,98,239],"associative":[69],"memory.":[70],"However,":[71],"not":[74],"been":[75],"scaled":[76],"LLM":[78],"pretraining":[79,188,212],"due":[80],"and":[83,107,134,171,179,182,202,214],"numerical":[84,103],"stability":[85],"concerns.":[86],"We":[87,118,138,174,209],"introduce":[88],"Parallax,":[89],"parameterized":[91],"that":[95,113,143,191],"scalable":[97],"LLMs.":[99],"Parallax":[100,120,176],"eliminates":[101],"solver":[104],"learns":[108],"extra":[110],"query-like":[111],"projector":[112],"probes":[114],"KV":[116],"covariance.":[117],"place":[119],"within":[121],"family":[123],"mechanisms":[126,241],"connected":[127],"by":[128],"bandwidth,":[130],"probe":[132],"construction":[133],"affine":[136],"structure.":[137],"propose":[139],"hardware-aware":[141],"algorithm":[142],"increases":[144],"arithmetic":[146],"intensity":[147],"over":[148],"FlashAttention,":[149],"shifting":[150],"into":[152],"more":[154],"compute":[155],"bound":[156],"regime.":[157],"Our":[158],"prototype":[159],"decode":[160],"kernel":[161],"matches":[162],"or":[163],"outperforms":[164],"FlashAttention":[165],"2/3":[166],"across":[167],"diverse":[168],"batch":[169],"sizes":[170],"context":[172],"lengths.":[173],"pretrain":[175],"at":[177],"0.6B":[178],"1.7B":[180],"scales":[181],"find":[183],"consistent":[184],"perplexity":[185],"improvements":[186],"throughout":[187],"with":[189],"gains":[190],"transfer":[192],"downstream":[194],"benchmarks.":[195],"The":[196],"advantage":[197],"persists":[198],"under":[199],"both":[200],"parameter-matched":[201],"compute-matched":[203],"controls,":[204],"demonstrating":[205],"Pareto":[207],"improvement.":[208],"perform":[210],"careful":[211],"ablations":[213],"identify":[215],"novel":[217],"phenomenon":[218],"whereby":[219],"Muon":[220],"unlocks":[221],"capacity":[223],"Parallax.":[225],"To":[226],"our":[227],"knowledge,":[228],"this":[229],"first":[232],"empirical":[233],"demonstration":[234],"strong":[236],"architecture-optimizer":[237],"codesign":[238],"architecture":[244],"literature.":[246]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-05-30T00:00:00"}