{"id":"https://openalex.org/W7141022825","doi":"https://doi.org/10.48550/arxiv.2603.24812","title":"Numerical Superoptimization for Library Learning","display_name":"Numerical Superoptimization for Library Learning","publication_year":2026,"publication_date":"2026-03-25","ids":{"openalex":"https://openalex.org/W7141022825","doi":"https://doi.org/10.48550/arxiv.2603.24812"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2603.24812","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.24812","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.2603.24812","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5114635287","display_name":"Jonas Regehr","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Regehr, Jonas","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5119926441","display_name":"Mitch Briles","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Briles, Mitch","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5042187082","display_name":"Zachary Tatlock","orcid":"https://orcid.org/0000-0002-4731-0124"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tatlock, Zachary","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5130749382","display_name":"Pavel Panchekha","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Panchekha, Pavel","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/T11697","display_name":"Numerical Methods and Algorithms","score":0.49630001187324524,"subfield":{"id":"https://openalex.org/subfields/1703","display_name":"Computational Theory and Mathematics"},"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/T11697","display_name":"Numerical Methods and Algorithms","score":0.49630001187324524,"subfield":{"id":"https://openalex.org/subfields/1703","display_name":"Computational Theory and Mathematics"},"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/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.19979999959468842,"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":0.05640000104904175,"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/counterfactual-thinking","display_name":"Counterfactual thinking","score":0.671500027179718},{"id":"https://openalex.org/keywords/implementation","display_name":"Implementation","score":0.6255999803543091},{"id":"https://openalex.org/keywords/workload","display_name":"Workload","score":0.5787000060081482},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.539900004863739},{"id":"https://openalex.org/keywords/set","display_name":"Set (abstract data type)","score":0.4966000020503998},{"id":"https://openalex.org/keywords/equivalence","display_name":"Equivalence (formal languages)","score":0.4625999927520752},{"id":"https://openalex.org/keywords/software","display_name":"Software","score":0.43700000643730164}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7350999712944031},{"id":"https://openalex.org/C108650721","wikidata":"https://www.wikidata.org/wiki/Q1783253","display_name":"Counterfactual thinking","level":2,"score":0.671500027179718},{"id":"https://openalex.org/C26713055","wikidata":"https://www.wikidata.org/wiki/Q245962","display_name":"Implementation","level":2,"score":0.6255999803543091},{"id":"https://openalex.org/C2778476105","wikidata":"https://www.wikidata.org/wiki/Q628539","display_name":"Workload","level":2,"score":0.5787000060081482},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.539900004863739},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.4966000020503998},{"id":"https://openalex.org/C2780069185","wikidata":"https://www.wikidata.org/wiki/Q7977945","display_name":"Equivalence (formal languages)","level":2,"score":0.4625999927520752},{"id":"https://openalex.org/C2777904410","wikidata":"https://www.wikidata.org/wiki/Q7397","display_name":"Software","level":2,"score":0.43700000643730164},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.4325000047683716},{"id":"https://openalex.org/C74193536","wikidata":"https://www.wikidata.org/wiki/Q574844","display_name":"Kernel (algebra)","level":2,"score":0.42559999227523804},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.3504999876022339},{"id":"https://openalex.org/C165696696","wikidata":"https://www.wikidata.org/wiki/Q11287","display_name":"Exploit","level":2,"score":0.34850001335144043},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3418000042438507},{"id":"https://openalex.org/C128099668","wikidata":"https://www.wikidata.org/wiki/Q573952","display_name":"Lazy evaluation","level":3,"score":0.3278000056743622},{"id":"https://openalex.org/C125583679","wikidata":"https://www.wikidata.org/wiki/Q755673","display_name":"Search algorithm","level":2,"score":0.3190000057220459},{"id":"https://openalex.org/C48753275","wikidata":"https://www.wikidata.org/wiki/Q11216","display_name":"Numerical analysis","level":2,"score":0.3181999921798706},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.31610000133514404},{"id":"https://openalex.org/C162319229","wikidata":"https://www.wikidata.org/wiki/Q175263","display_name":"Data structure","level":2,"score":0.3131999969482422},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.2806999981403351},{"id":"https://openalex.org/C58328972","wikidata":"https://www.wikidata.org/wiki/Q184609","display_name":"Expert system","level":2,"score":0.2653000056743622}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2603.24812","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.24812","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.2603.24812","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.24812","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":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Numerical":[0],"software":[1],"depends":[2],"on":[3],"fast,":[4],"accurate":[5],"implementations":[6,64],"of":[7,25,55],"mathematical":[8,60],"primitives":[9,37,61,138,166,216],"like":[10],"sin,":[11],"exp,":[12],"and":[13,69,99,102,150,160,194,228],"log.":[14],"Modern":[15],"superoptimizers":[16,77],"can":[17,97,106,168],"optimize":[18],"floating-point":[19,56],"kernels":[20],"against":[21],"a":[22,29,53,117,131],"given":[23,52,118],"set":[24],"such":[26],"primitives,":[27,93],"but":[28],"more":[30],"fundamental":[31],"question":[32],"remains":[33],"open:":[34],"which":[35,125],"new":[36],"are":[38],"worth":[39],"implementing":[40],"in":[41,200,217],"the":[42,59,80,112,127,140],"first":[43],"place?":[44],"We":[45,122,207],"formulate":[46],"this":[47,84],"as":[48,130],"numerical":[49,76,132,174],"library":[50,133],"learning:":[51],"workload":[54,113],"kernels,":[57],"identify":[58],"whose":[62],"expert":[63,182],"would":[65,114],"most":[66],"improve":[67,115],"speed":[68,185],"accuracy.":[70],"Our":[71],"key":[72],"insight":[73],"is":[74,178],"that":[75,167,213],"already":[78],"have":[79],"machinery":[81],"well-suited":[82],"to":[83,90,189,205,235],"problem.":[85],"Their":[86],"search":[87,143],"procedures":[88,96],"happen":[89],"enumerate":[91],"candidate":[92,137],"their":[94,103],"equivalence":[95],"generalize":[98],"deduplicate":[100],"candidates,":[101],"cost":[104],"models":[105],"estimate":[107],"counterfactual":[108,148],"utility:":[109],"how":[110],"much":[111],"if":[116],"primitive":[119],"were":[120],"available.":[121],"present":[123],"GrowLibm,":[124],"repurposes":[126],"Herbie":[128,177],"superoptimizer":[129],"learner.":[134],"GrowLibm":[135,162],"mines":[136],"from":[139,203],"superoptimizer's":[141],"intermediate":[142],"results,":[144],"ranks":[145],"them":[146],"by":[147,187,233],"utility,":[149],"prunes":[151],"redundant":[152],"candidates.":[153],"Across":[154],"three":[155],"scientific":[156],"applications":[157],"(PROJ,":[158],"CoolProp,":[159],"Basilisk),":[161],"identifies":[163],"compact,":[164],"reusable":[165],"be":[169],"implemented":[170],"effectively":[171],"using":[172],"standard":[173],"techniques.":[175],"When":[176],"extended":[179],"with":[180],"these":[181],"implementations,":[183],"kernel":[184],"improves":[186],"up":[188,234],"2.2x":[190],"at":[191],"fixed":[192],"accuracy,":[193],"maximum":[195],"achievable":[196],"accuracy":[197],"also":[198,208],"improves,":[199],"one":[201],"case":[202],"56.0%":[204],"93.5%.":[206],"prototype":[209],"an":[210],"LLVM":[211],"matcher":[212],"recognizes":[214],"learned":[215],"optimized":[218],"IR,":[219],"recovering":[220],"26":[221],"replacement":[222],"sites":[223],"across":[224],"five":[225],"PROJ":[226],"projections":[227],"improving":[229],"end-to-end":[230],"application":[231],"performance":[232],"5%.":[236]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-03-28T00:00:00"}