{"id":"https://openalex.org/W7139914420","doi":"https://doi.org/10.1016/j.procs.2026.01.048","title":"TurkColBERT: A Benchmark of Dense and Late-Interaction Models for Turkish Information Retrieval","display_name":"TurkColBERT: A Benchmark of Dense and Late-Interaction Models for Turkish Information Retrieval","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7139914420","doi":"https://doi.org/10.1016/j.procs.2026.01.048"},"language":"en","primary_location":{"id":"doi:10.1016/j.procs.2026.01.048","is_oa":true,"landing_page_url":"https://doi.org/10.1016/j.procs.2026.01.048","pdf_url":null,"source":{"id":"https://openalex.org/S120348307","display_name":"Procedia Computer Science","issn_l":"1877-0509","issn":["1877-0509"],"is_oa":true,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310320990","host_organization_name":"Elsevier BV","host_organization_lineage":["https://openalex.org/P4310320990"],"host_organization_lineage_names":["Elsevier BV"],"type":"journal"},"license":"cc-by-nc-nd","license_id":"https://openalex.org/licenses/cc-by-nc-nd","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Procedia Computer Science","raw_type":"journal-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"diamond","oa_url":"https://doi.org/10.1016/j.procs.2026.01.048","any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5119399852","display_name":"\u00d6zay Ezerceli","orcid":"https://orcid.org/0000-0002-7877-7528"},"institutions":[{"id":"https://openalex.org/I4210103784","display_name":"Small and Medium Enterprises Development Organization","ror":"https://ror.org/025rbqk82","country_code":"TR","type":"government","lineage":["https://openalex.org/I4210103784"]}],"countries":["TR"],"is_corresponding":true,"raw_author_name":"\u00d6zay Ezerceli","raw_affiliation_strings":["NewMind AI, Maslak Mahallesi, Sar\u0131yer, Istanbul, 34398, T\u00fcrkiye"],"affiliations":[{"raw_affiliation_string":"NewMind AI, Maslak Mahallesi, Sar\u0131yer, Istanbul, 34398, T\u00fcrkiye","institution_ids":["https://openalex.org/I4210103784"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5120193201","display_name":"Mahmoud El Hussieni","orcid":null},"institutions":[{"id":"https://openalex.org/I4210103784","display_name":"Small and Medium Enterprises Development Organization","ror":"https://ror.org/025rbqk82","country_code":"TR","type":"government","lineage":["https://openalex.org/I4210103784"]}],"countries":["TR"],"is_corresponding":false,"raw_author_name":"Mahmoud El Hussieni","raw_affiliation_strings":["NewMind AI, Maslak Mahallesi, Sar\u0131yer, Istanbul, 34398, T\u00fcrkiye"],"affiliations":[{"raw_affiliation_string":"NewMind AI, Maslak Mahallesi, Sar\u0131yer, Istanbul, 34398, T\u00fcrkiye","institution_ids":["https://openalex.org/I4210103784"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5126432634","display_name":"Selva Ta\u015f","orcid":null},"institutions":[{"id":"https://openalex.org/I4210103784","display_name":"Small and Medium Enterprises Development Organization","ror":"https://ror.org/025rbqk82","country_code":"TR","type":"government","lineage":["https://openalex.org/I4210103784"]}],"countries":["TR"],"is_corresponding":false,"raw_author_name":"Selva Ta\u015f","raw_affiliation_strings":["NewMind AI, Maslak Mahallesi, Sar\u0131yer, Istanbul, 34398, T\u00fcrkiye"],"affiliations":[{"raw_affiliation_string":"NewMind AI, Maslak Mahallesi, Sar\u0131yer, Istanbul, 34398, T\u00fcrkiye","institution_ids":["https://openalex.org/I4210103784"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5119399854","display_name":"Reyhan Bayraktar","orcid":null},"institutions":[{"id":"https://openalex.org/I4210103784","display_name":"Small and Medium Enterprises Development Organization","ror":"https://ror.org/025rbqk82","country_code":"TR","type":"government","lineage":["https://openalex.org/I4210103784"]}],"countries":["TR"],"is_corresponding":false,"raw_author_name":"Reyhan Bayraktar","raw_affiliation_strings":["NewMind AI, Maslak Mahallesi, Sar\u0131yer, Istanbul, 34398, T\u00fcrkiye"],"affiliations":[{"raw_affiliation_string":"NewMind AI, Maslak Mahallesi, Sar\u0131yer, Istanbul, 34398, T\u00fcrkiye","institution_ids":["https://openalex.org/I4210103784"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5033525022","display_name":"Fatma Terzioglu","orcid":"https://orcid.org/0000-0001-8448-672X"},"institutions":[{"id":"https://openalex.org/I4210103784","display_name":"Small and Medium Enterprises Development Organization","ror":"https://ror.org/025rbqk82","country_code":"TR","type":"government","lineage":["https://openalex.org/I4210103784"]}],"countries":["TR"],"is_corresponding":false,"raw_author_name":"Fatma Bet\u00fcl Terzio\u011flu","raw_affiliation_strings":["NewMind AI, Maslak Mahallesi, Sar\u0131yer, Istanbul, 34398, T\u00fcrkiye"],"affiliations":[{"raw_affiliation_string":"NewMind AI, Maslak Mahallesi, Sar\u0131yer, Istanbul, 34398, T\u00fcrkiye","institution_ids":["https://openalex.org/I4210103784"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5120424512","display_name":"Yusuf \u00c7elebi","orcid":null},"institutions":[{"id":"https://openalex.org/I4210103784","display_name":"Small and Medium Enterprises Development Organization","ror":"https://ror.org/025rbqk82","country_code":"TR","type":"government","lineage":["https://openalex.org/I4210103784"]}],"countries":["TR"],"is_corresponding":false,"raw_author_name":"Yusuf \u00c7elebi","raw_affiliation_strings":["NewMind AI, Maslak Mahallesi, Sar\u0131yer, Istanbul, 34398, T\u00fcrkiye"],"affiliations":[{"raw_affiliation_string":"NewMind AI, Maslak Mahallesi, Sar\u0131yer, Istanbul, 34398, T\u00fcrkiye","institution_ids":["https://openalex.org/I4210103784"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5130219645","display_name":"Ya\u011f\u0131z Asker","orcid":null},"institutions":[{"id":"https://openalex.org/I4210103784","display_name":"Small and Medium Enterprises Development Organization","ror":"https://ror.org/025rbqk82","country_code":"TR","type":"government","lineage":["https://openalex.org/I4210103784"]}],"countries":["TR"],"is_corresponding":false,"raw_author_name":"Ya\u011f\u0131z Asker","raw_affiliation_strings":["NewMind AI, Maslak Mahallesi, Sar\u0131yer, Istanbul, 34398, T\u00fcrkiye"],"affiliations":[{"raw_affiliation_string":"NewMind AI, Maslak Mahallesi, Sar\u0131yer, Istanbul, 34398, T\u00fcrkiye","institution_ids":["https://openalex.org/I4210103784"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":7,"corresponding_author_ids":["https://openalex.org/A5119399852"],"corresponding_institution_ids":["https://openalex.org/I4210103784"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.94081399,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":null,"biblio":{"volume":"275","issue":null,"first_page":"393","last_page":"400"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10286","display_name":"Information Retrieval and Search Behavior","score":0.7567999958992004,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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/T10286","display_name":"Information Retrieval and Search Behavior","score":0.7567999958992004,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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":0.027699999511241913,"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/T11550","display_name":"Text and Document Classification Technologies","score":0.025299999862909317,"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/benchmark","display_name":"Benchmark (surveying)","score":0.7544000148773193},{"id":"https://openalex.org/keywords/turkish","display_name":"Turkish","score":0.6675000190734863},{"id":"https://openalex.org/keywords/query-expansion","display_name":"Query expansion","score":0.32679998874664307},{"id":"https://openalex.org/keywords/data-retrieval","display_name":"Data retrieval","score":0.32170000672340393},{"id":"https://openalex.org/keywords/vector-space-model","display_name":"Vector space model","score":0.302700012922287}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.9128000140190125},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.7544000148773193},{"id":"https://openalex.org/C2781121862","wikidata":"https://www.wikidata.org/wiki/Q256","display_name":"Turkish","level":2,"score":0.6675000190734863},{"id":"https://openalex.org/C23123220","wikidata":"https://www.wikidata.org/wiki/Q816826","display_name":"Information retrieval","level":1,"score":0.6093999743461609},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4666999876499176},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.4172999858856201},{"id":"https://openalex.org/C99016210","wikidata":"https://www.wikidata.org/wiki/Q5488129","display_name":"Query expansion","level":2,"score":0.32679998874664307},{"id":"https://openalex.org/C551230270","wikidata":"https://www.wikidata.org/wiki/Q4368942","display_name":"Data retrieval","level":2,"score":0.32170000672340393},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.32089999318122864},{"id":"https://openalex.org/C89686163","wikidata":"https://www.wikidata.org/wiki/Q1187982","display_name":"Vector space model","level":2,"score":0.302700012922287},{"id":"https://openalex.org/C161156560","wikidata":"https://www.wikidata.org/wiki/Q1638872","display_name":"Document retrieval","level":2,"score":0.2851000130176544},{"id":"https://openalex.org/C2983174267","wikidata":"https://www.wikidata.org/wiki/Q3775098","display_name":"Video retrieval","level":2,"score":0.28049999475479126},{"id":"https://openalex.org/C90288658","wikidata":"https://www.wikidata.org/wiki/Q3318149","display_name":"Human\u2013computer information retrieval","level":3,"score":0.27900001406669617},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.2655999958515167}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1016/j.procs.2026.01.048","is_oa":true,"landing_page_url":"https://doi.org/10.1016/j.procs.2026.01.048","pdf_url":null,"source":{"id":"https://openalex.org/S120348307","display_name":"Procedia Computer Science","issn_l":"1877-0509","issn":["1877-0509"],"is_oa":true,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310320990","host_organization_name":"Elsevier BV","host_organization_lineage":["https://openalex.org/P4310320990"],"host_organization_lineage_names":["Elsevier BV"],"type":"journal"},"license":"cc-by-nc-nd","license_id":"https://openalex.org/licenses/cc-by-nc-nd","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Procedia Computer Science","raw_type":"journal-article"}],"best_oa_location":{"id":"doi:10.1016/j.procs.2026.01.048","is_oa":true,"landing_page_url":"https://doi.org/10.1016/j.procs.2026.01.048","pdf_url":null,"source":{"id":"https://openalex.org/S120348307","display_name":"Procedia Computer Science","issn_l":"1877-0509","issn":["1877-0509"],"is_oa":true,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310320990","host_organization_name":"Elsevier BV","host_organization_lineage":["https://openalex.org/P4310320990"],"host_organization_lineage_names":["Elsevier BV"],"type":"journal"},"license":"cc-by-nc-nd","license_id":"https://openalex.org/licenses/cc-by-nc-nd","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Procedia Computer Science","raw_type":"journal-article"},"sustainable_development_goals":[{"score":0.5093197226524353,"id":"https://metadata.un.org/sdg/4","display_name":"Quality Education"}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":11,"referenced_works":["https://openalex.org/W2970641574","https://openalex.org/W3021397474","https://openalex.org/W3035390927","https://openalex.org/W3099700870","https://openalex.org/W3154280800","https://openalex.org/W3185274669","https://openalex.org/W3217305727","https://openalex.org/W4306317212","https://openalex.org/W4404783220","https://openalex.org/W4415796011","https://openalex.org/W4416018085"],"related_works":[],"abstract_inverted_index":{"Neural":[0],"information":[1,65],"retrieval":[2,48,267,272],"systems":[3,211],"have":[4,41],"achieved":[5],"remarkable":[6],"success":[7],"for":[8,37,46],"high-resource":[9],"languages,":[10],"yet":[11,142],"their":[12],"application":[13],"to":[14,167,239,250,280],"morphologically":[15],"complex,":[16],"lower-resource":[17],"languages":[18],"like":[19,216],"Turkish":[20,28,47,64,85,110,242,266],"remains":[21],"underexplored.":[22],"While":[23],"dense":[24,58,140,161],"bi-encoders":[25],"dominate":[26],"current":[27],"IR":[29,243],"systems,":[30],"late-interaction":[31,61,157],"models":[32,62,106,121,158,215],"which":[33,260],"preserve":[34],"token-level":[35],"representations":[36],"fine-grained":[38],"semantic":[39],"matching":[40],"not":[42,262],"been":[43],"systematically":[44],"evaluated":[45,104],"tasks.":[49,174],"We":[50,67,103,228],"present":[51],"TurkColBERT,":[52],"the":[53,136,193],"first":[54],"comprehensive":[55],"benchmark":[56],"comparing":[57],"encoders":[59,82],"and":[60,79,87,116,236,257],"on":[63,84,100,172,188,283],"retrieval.":[66],"introduce":[68],"a":[69,200],"two-stage":[70],"adaptation":[71],"pipeline:":[72],"first,":[73],"we":[74,91,178],"fine-tune":[75],"english-based":[76],"(ettin,":[77],"bert-hash)":[78],"modern":[80],"multilingual":[81],"(mmBERT)":[83],"NLI":[86],"STS":[88],"tasks;":[89],"second,":[90],"convert":[92],"them":[93],"into":[94],"ColBERT-style":[95],"retrievers":[96],"using":[97,226],"PyLate,":[98],"trained":[99],"MS":[101],"MARCO-TR.":[102],"10":[105],"across":[107],"five":[108],"diverse":[109],"BEIR":[111],"datasets":[112,253],"spanning":[113],"scientific,":[114],"financial,":[115],"argumentative":[117],"domains.":[118],"Notably,":[119],"these":[120],"show":[122,153],"exceptional":[123],"parameter":[124,128,138],"efficiency:":[125],"our":[126,246],"1.0M":[127],"colbert-hash-nano-tr":[129],"model":[130,232],"is":[131,248,278],"600":[132],"times":[133,220],"smaller":[134,156],"than":[135,192],"600M":[137],"turkish-e5-large":[139],"encoder,":[141],"retains":[143],"over":[144],"71%":[145],"of":[146],"its":[147],"average":[148],"mAP":[149,171],"performance.":[150],"Our":[151,182],"results":[152],"that":[154,185,208],"3-5x":[155],"significantly":[159],"outperform":[160],"encoders,":[162],"with":[163,214],"ColmmBERT-base-TR":[164,217],"achieving":[165,199,218],"up":[166],"+13.8%":[168],"improvement":[169,203],"in":[170,204],"domain-specific":[173],"To":[175],"assess":[176,281],"production-readiness,":[177],"benchmarked":[179],"indexing":[180],"algorithms.":[181],"analysis":[183],"shows":[184,273],"MUVERA+Rerank":[186],"is,":[187],"average,":[189],"3.33x":[190],"faster":[191],"standard":[194],"PLAID":[195],"index":[196],"while":[197,270],"also":[198],"+1.7%":[201],"relative":[202],"mAP.":[205],"This":[206],"demonstrates":[207],"efficient,":[209],"low-latency":[210],"are":[212],"practical,":[213],"query":[219],"as":[221,223],"low":[222],"0.54":[224],"ms":[225],"MUVERA.":[227],"publicly":[229],"release":[230],"all":[231],"checkpoints,":[233],"training":[234],"configurations,":[235],"evaluation":[237,277],"scripts":[238],"facilitate":[240],"future":[241],"research.":[244],"However,":[245],"study":[247],"limited":[249],"moderately":[251],"sized":[252],"(\u2264":[254],"50K":[255],"documents)":[256],"translated":[258],"benchmarks,":[259],"may":[261],"fully":[263],"reflect":[264],"real-world":[265],"scenarios.":[268],"Additionally,":[269],"MUVERA-based":[271],"promising":[274],"latency,":[275],"further":[276],"needed":[279],"scalability":[282],"large-scale":[284],"production":[285],"systems.":[286]},"counts_by_year":[],"updated_date":"2026-03-22T06:25:25.174409","created_date":"2026-03-21T00:00:00"}