{"id":"https://openalex.org/W7135229600","doi":"https://doi.org/10.48550/arxiv.2603.11622","title":"Sema: A High-performance System for LLM-based Semantic Query Processing","display_name":"Sema: A High-performance System for LLM-based Semantic Query Processing","publication_year":2026,"publication_date":"2026-03-12","ids":{"openalex":"https://openalex.org/W7135229600","doi":"https://doi.org/10.48550/arxiv.2603.11622"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2603.11622","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.11622","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,"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":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.2603.11622","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5128975922","display_name":"Kangkang Qi","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Qi, Kangkang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129069558","display_name":"Dongyang Xie","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xie, Dongyang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129094662","display_name":"Wenbo Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Wenbo","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128941277","display_name":"Hao Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Hao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128933895","display_name":"Yuanyuan Zhu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhu, Yuanyuan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129069580","display_name":"Jeffrey Xu Yu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yu, Jeffrey Xu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5128943220","display_name":"Kangfei Zhao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhao, Kangfei","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/T10317","display_name":"Advanced Database Systems and Queries","score":0.1265999972820282,"subfield":{"id":"https://openalex.org/subfields/1705","display_name":"Computer Networks and Communications"},"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/T10317","display_name":"Advanced Database Systems and Queries","score":0.1265999972820282,"subfield":{"id":"https://openalex.org/subfields/1705","display_name":"Computer Networks and Communications"},"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":0.1234000027179718,"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/T10215","display_name":"Semantic Web and Ontologies","score":0.12120000272989273,"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/query-language","display_name":"Query language","score":0.6064000129699707},{"id":"https://openalex.org/keywords/query-optimization","display_name":"Query optimization","score":0.5296000242233276},{"id":"https://openalex.org/keywords/rdf-query-language","display_name":"RDF query language","score":0.5120000243186951},{"id":"https://openalex.org/keywords/semantic-query","display_name":"Semantic query","score":0.49950000643730164},{"id":"https://openalex.org/keywords/query-plan","display_name":"Query plan","score":0.4629000127315521},{"id":"https://openalex.org/keywords/semantic-computing","display_name":"Semantic computing","score":0.445499986410141},{"id":"https://openalex.org/keywords/sql","display_name":"SQL","score":0.4302999973297119},{"id":"https://openalex.org/keywords/datalog","display_name":"Datalog","score":0.4083999991416931},{"id":"https://openalex.org/keywords/data-manipulation-language","display_name":"Data manipulation language","score":0.4036000072956085},{"id":"https://openalex.org/keywords/sparql","display_name":"SPARQL","score":0.37720000743865967}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8960999846458435},{"id":"https://openalex.org/C192028432","wikidata":"https://www.wikidata.org/wiki/Q845739","display_name":"Query language","level":2,"score":0.6064000129699707},{"id":"https://openalex.org/C157692150","wikidata":"https://www.wikidata.org/wiki/Q2919848","display_name":"Query optimization","level":2,"score":0.5296000242233276},{"id":"https://openalex.org/C96956885","wikidata":"https://www.wikidata.org/wiki/Q6138701","display_name":"RDF query language","level":5,"score":0.5120000243186951},{"id":"https://openalex.org/C2778816267","wikidata":"https://www.wikidata.org/wiki/Q21015578","display_name":"Semantic query","level":4,"score":0.49950000643730164},{"id":"https://openalex.org/C2779729312","wikidata":"https://www.wikidata.org/wiki/Q784232","display_name":"Query plan","level":5,"score":0.4629000127315521},{"id":"https://openalex.org/C511149849","wikidata":"https://www.wikidata.org/wiki/Q7449051","display_name":"Semantic computing","level":3,"score":0.445499986410141},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.430400013923645},{"id":"https://openalex.org/C510870499","wikidata":"https://www.wikidata.org/wiki/Q47607","display_name":"SQL","level":2,"score":0.4302999973297119},{"id":"https://openalex.org/C148230440","wikidata":"https://www.wikidata.org/wiki/Q1172264","display_name":"Datalog","level":2,"score":0.4083999991416931},{"id":"https://openalex.org/C56288433","wikidata":"https://www.wikidata.org/wiki/Q58673","display_name":"Data manipulation language","level":2,"score":0.4036000072956085},{"id":"https://openalex.org/C41009113","wikidata":"https://www.wikidata.org/wiki/Q54871","display_name":"SPARQL","level":4,"score":0.37720000743865967},{"id":"https://openalex.org/C23123220","wikidata":"https://www.wikidata.org/wiki/Q816826","display_name":"Information retrieval","level":1,"score":0.3709000051021576},{"id":"https://openalex.org/C147497476","wikidata":"https://www.wikidata.org/wiki/Q54872","display_name":"RDF","level":3,"score":0.3698999881744385},{"id":"https://openalex.org/C90312973","wikidata":"https://www.wikidata.org/wiki/Q7449052","display_name":"Semantic data model","level":2,"score":0.36640000343322754},{"id":"https://openalex.org/C195324797","wikidata":"https://www.wikidata.org/wiki/Q33742","display_name":"Natural language","level":2,"score":0.3573000133037567},{"id":"https://openalex.org/C2129575","wikidata":"https://www.wikidata.org/wiki/Q54837","display_name":"Semantic Web","level":2,"score":0.3425000011920929},{"id":"https://openalex.org/C192939062","wikidata":"https://www.wikidata.org/wiki/Q104840822","display_name":"Sargable","level":4,"score":0.3402999937534332},{"id":"https://openalex.org/C5655090","wikidata":"https://www.wikidata.org/wiki/Q192588","display_name":"Relational database","level":2,"score":0.33059999346733093},{"id":"https://openalex.org/C98199447","wikidata":"https://www.wikidata.org/wiki/Q2445044","display_name":"Materialized view","level":4,"score":0.3197000026702881},{"id":"https://openalex.org/C194222762","wikidata":"https://www.wikidata.org/wiki/Q114486","display_name":"Query by Example","level":4,"score":0.3188999891281128},{"id":"https://openalex.org/C61114434","wikidata":"https://www.wikidata.org/wiki/Q7144649","display_name":"Path expression","level":3,"score":0.31610000133514404},{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.313400000333786},{"id":"https://openalex.org/C102993220","wikidata":"https://www.wikidata.org/wiki/Q387196","display_name":"Description logic","level":2,"score":0.31119999289512634},{"id":"https://openalex.org/C148792806","wikidata":"https://www.wikidata.org/wiki/Q7449046","display_name":"Semantic analytics","level":4,"score":0.31049999594688416},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.3046000003814697},{"id":"https://openalex.org/C179531526","wikidata":"https://www.wikidata.org/wiki/Q595637","display_name":"Language Integrated Query","level":5,"score":0.2996000051498413},{"id":"https://openalex.org/C174252522","wikidata":"https://www.wikidata.org/wiki/Q3816772","display_name":"Natural language user interface","level":3,"score":0.29249998927116394},{"id":"https://openalex.org/C99016210","wikidata":"https://www.wikidata.org/wiki/Q5488129","display_name":"Query expansion","level":2,"score":0.29010000824928284},{"id":"https://openalex.org/C37926939","wikidata":"https://www.wikidata.org/wiki/Q7449061","display_name":"Semantic equivalence","level":4,"score":0.2890999913215637},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.2865999937057495},{"id":"https://openalex.org/C202708506","wikidata":"https://www.wikidata.org/wiki/Q7449050","display_name":"Semantic compression","level":5,"score":0.27559998631477356},{"id":"https://openalex.org/C71695816","wikidata":"https://www.wikidata.org/wiki/Q3478367","display_name":"Semantic Web Rule Language","level":5,"score":0.27300000190734863},{"id":"https://openalex.org/C81721847","wikidata":"https://www.wikidata.org/wiki/Q163468","display_name":"Prolog","level":2,"score":0.2685999870300293},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.2628999948501587},{"id":"https://openalex.org/C128838566","wikidata":"https://www.wikidata.org/wiki/Q275603","display_name":"Logic programming","level":2,"score":0.2563999891281128},{"id":"https://openalex.org/C6881194","wikidata":"https://www.wikidata.org/wiki/Q7449091","display_name":"Semantic technology","level":4,"score":0.2549999952316284},{"id":"https://openalex.org/C2776187449","wikidata":"https://www.wikidata.org/wiki/Q1513879","display_name":"Natural language generation","level":3,"score":0.25450000166893005},{"id":"https://openalex.org/C184337299","wikidata":"https://www.wikidata.org/wiki/Q1437428","display_name":"Semantics (computer science)","level":2,"score":0.2515000104904175}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2603.11622","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.11622","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,"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":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2603.11622","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.11622","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,"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":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"sustainable_development_goals":[{"score":0.6415045261383057,"display_name":"Industry, innovation and infrastructure","id":"https://metadata.un.org/sdg/9"}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"The":[0,59],"integration":[1],"of":[2,37,75,142],"Large":[3],"Language":[4],"Models":[5],"(LLMs)":[6],"into":[7,127],"data":[8],"analytics":[9],"has":[10],"unlocked":[11],"powerful":[12],"capabilities":[13],"for":[14,87],"reasoning":[15],"over":[16],"bulk":[17],"structured":[18],"and":[19,53,72,82,134,148,170,184,199],"unstructured":[20],"data.":[21],"However,":[22],"existing":[23],"systems":[24,214],"typically":[25],"rely":[26],"on":[27,103,192],"either":[28],"DataFrame":[29],"primitives,":[30],"which":[31,45],"lack":[32],"the":[33,50,69,137,140],"efficient":[34],"execution":[35,179],"infrastructure":[36],"modern":[38],"DBMSs,":[39],"or":[40],"SQL":[41,129],"User-Defined":[42],"Functions":[43],"(UDFs),":[44],"isolate":[46],"semantic":[47,61,99,108,153,168,194],"logic":[48],"from":[49,152],"query":[51,100],"optimizer":[52,141],"burden":[54],"users":[55,121],"with":[56],"implementation":[57],"complexities.":[58],"LLM-powered":[60,107],"operators":[62,109],"also":[63],"bring":[64],"new":[65],"challenges":[66],"due":[67],"to":[68,163],"high":[70],"cost":[71,83],"non-deterministic":[73],"nature":[74],"LLM":[76],"invocation,":[77],"where":[78],"conventional":[79],"optimization":[80,133],"rules":[81],"models":[84],"are":[85],"inapplicable":[86],"their":[88],"optimization.":[89],"To":[90],"bridge":[91],"these":[92],"gaps,":[93],"we":[94],"present":[95],"Sema,":[96],"a":[97,116,177],"high-performance":[98],"engine":[101],"built":[102],"DuckDB":[104],"that":[105,119,205],"treats":[106],"as":[110],"first-class":[111],"citizens.":[112],"Sema":[113,143,157,191,206],"introduces":[114],"SemaSQL,":[115],"declarative":[117],"dialect":[118],"allows":[120],"seamlessly":[122],"inject":[123],"natural":[124,145],"language":[125,146],"expressions":[126,147],"standard":[128],"clauses,":[130],"enabling":[131],"end-to-end":[132],"execution.":[135],"At":[136,155],"logical":[138],"level,":[139],"compresses":[144],"deduces":[149],"relational":[150],"constraints":[151],"operators.":[154],"runtime,":[156],"employs":[158],"Adaptive":[159],"Query":[160],"Execution":[161],"(AQE)":[162],"dynamically":[164],"reorder":[165],"operators,":[166],"fuse":[167],"operations,":[169],"apply":[171],"prompt":[172],"batching.":[173],"This":[174],"approach":[175],"seeks":[176],"Pareto-optimal":[178],"path":[180],"balancing":[181],"token":[182],"consumption":[183],"latency":[185],"under":[186],"accuracy":[187],"constraints.":[188],"We":[189],"evaluate":[190],"20":[193],"queries":[195],"across":[196],"classification,":[197],"summarization,":[198],"extraction":[200],"tasks.":[201],"Experimental":[202],"results":[203],"demonstrate":[204],"achieves":[207],"$2-10":[208],"\\times$":[209],"speedup":[210],"against":[211],"three":[212],"baseline":[213],"while":[215],"achieving":[216],"competitive":[217],"result":[218],"quality.":[219]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2026-03-14T00:00:00"}