{"id":"https://openalex.org/W7165377867","doi":"https://doi.org/10.48550/arxiv.2606.20089","title":"IHUBERT: Vector-Based Semantic Deduplication and Domain-Balanced Pretraining for Persian Resources","display_name":"IHUBERT: Vector-Based Semantic Deduplication and Domain-Balanced Pretraining for Persian Resources","publication_year":2026,"publication_date":"2026-06-18","ids":{"openalex":"https://openalex.org/W7165377867","doi":"https://doi.org/10.48550/arxiv.2606.20089"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2606.20089","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.20089","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":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.2606.20089","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5053976236","display_name":"Arash Ghafouri","orcid":"https://orcid.org/0009-0005-9828-085X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ghafouri, Arash","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5046517153","display_name":"Mahdi Firouzmandi","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Firouzmandi, Mahdi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139015790","display_name":"Hossein Saberi","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Saberi, Hossein","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5139022286","display_name":"Mohammad Reza Hasani Ahangar","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ahangar, Mohammad Reza Hasani","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/T10028","display_name":"Topic Modeling","score":0.531499981880188,"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"}},"topics":[{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.531499981880188,"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":0.18140000104904175,"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/T13629","display_name":"Text Readability and Simplification","score":0.08640000224113464,"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/preprocessor","display_name":"Preprocessor","score":0.5157999992370605},{"id":"https://openalex.org/keywords/persian","display_name":"Persian","score":0.5142999887466431},{"id":"https://openalex.org/keywords/vocabulary","display_name":"Vocabulary","score":0.44359999895095825},{"id":"https://openalex.org/keywords/lexical-analysis","display_name":"Lexical analysis","score":0.40880000591278076},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.35370001196861267},{"id":"https://openalex.org/keywords/relationship-extraction","display_name":"Relationship extraction","score":0.3506999909877777},{"id":"https://openalex.org/keywords/pipeline","display_name":"Pipeline (software)","score":0.3221000134944916},{"id":"https://openalex.org/keywords/relation","display_name":"Relation (database)","score":0.3199000060558319},{"id":"https://openalex.org/keywords/prefix","display_name":"Prefix","score":0.30820000171661377}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8192999958992004},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.6934999823570251},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.6732000112533569},{"id":"https://openalex.org/C34736171","wikidata":"https://www.wikidata.org/wiki/Q918333","display_name":"Preprocessor","level":2,"score":0.5157999992370605},{"id":"https://openalex.org/C2776527531","wikidata":"https://www.wikidata.org/wiki/Q9168","display_name":"Persian","level":2,"score":0.5142999887466431},{"id":"https://openalex.org/C2777601683","wikidata":"https://www.wikidata.org/wiki/Q6499736","display_name":"Vocabulary","level":2,"score":0.44359999895095825},{"id":"https://openalex.org/C176982825","wikidata":"https://www.wikidata.org/wiki/Q835922","display_name":"Lexical analysis","level":2,"score":0.40880000591278076},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.35370001196861267},{"id":"https://openalex.org/C153604712","wikidata":"https://www.wikidata.org/wiki/Q7310755","display_name":"Relationship extraction","level":3,"score":0.3506999909877777},{"id":"https://openalex.org/C43521106","wikidata":"https://www.wikidata.org/wiki/Q2165493","display_name":"Pipeline (software)","level":2,"score":0.3221000134944916},{"id":"https://openalex.org/C25343380","wikidata":"https://www.wikidata.org/wiki/Q277521","display_name":"Relation (database)","level":2,"score":0.3199000060558319},{"id":"https://openalex.org/C23123220","wikidata":"https://www.wikidata.org/wiki/Q816826","display_name":"Information retrieval","level":1,"score":0.31790000200271606},{"id":"https://openalex.org/C141603448","wikidata":"https://www.wikidata.org/wiki/Q134830","display_name":"Prefix","level":2,"score":0.30820000171661377},{"id":"https://openalex.org/C184337299","wikidata":"https://www.wikidata.org/wiki/Q1437428","display_name":"Semantics (computer science)","level":2,"score":0.30820000171661377},{"id":"https://openalex.org/C191015642","wikidata":"https://www.wikidata.org/wiki/Q1132459","display_name":"Fragmentation (computing)","level":2,"score":0.3005000054836273},{"id":"https://openalex.org/C189430467","wikidata":"https://www.wikidata.org/wiki/Q7293293","display_name":"Ranking (information retrieval)","level":2,"score":0.2849999964237213},{"id":"https://openalex.org/C191399111","wikidata":"https://www.wikidata.org/wiki/Q64861","display_name":"Transitive relation","level":2,"score":0.2711000144481659},{"id":"https://openalex.org/C2780945871","wikidata":"https://www.wikidata.org/wiki/Q194274","display_name":"Backup","level":2,"score":0.2705000042915344},{"id":"https://openalex.org/C148524875","wikidata":"https://www.wikidata.org/wiki/Q6975395","display_name":"F1 score","level":2,"score":0.26649999618530273},{"id":"https://openalex.org/C118505674","wikidata":"https://www.wikidata.org/wiki/Q42586063","display_name":"Encoder","level":2,"score":0.2632000148296356},{"id":"https://openalex.org/C110615152","wikidata":"https://www.wikidata.org/wiki/Q1469824","display_name":"Controlled vocabulary","level":2,"score":0.26260000467300415},{"id":"https://openalex.org/C32587265","wikidata":"https://www.wikidata.org/wiki/Q1182260","display_name":"Data deduplication","level":2,"score":0.2621999979019165},{"id":"https://openalex.org/C2474386","wikidata":"https://www.wikidata.org/wiki/Q461183","display_name":"Text corpus","level":2,"score":0.25940001010894775},{"id":"https://openalex.org/C2779135771","wikidata":"https://www.wikidata.org/wiki/Q403574","display_name":"Named-entity recognition","level":3,"score":0.2522999942302704},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.2515000104904175}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2606.20089","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.20089","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":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.2606.20089","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.20089","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"sustainable_development_goals":[{"score":0.6743537187576294,"display_name":"Quality Education","id":"https://metadata.un.org/sdg/4"}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Persian":[0,31,103,113,223],"pretrained":[1],"language":[2,224],"models":[3],"(PLMs)":[4],"are":[5],"still":[6],"limited":[7],"by":[8,17],"the":[9,37,49,96,157,185,198],"scarcity":[10],"of":[11,48],"large-scale,":[12],"high-quality":[13],"pretraining":[14,98,200,230],"corpora":[15],"and":[16,23,59,72,76,86,105,126,151,155,166,231,237],"insufficient":[18],"evaluation":[19,233],"beyond":[20],"standard":[21],"classification":[22,236],"NER":[24,165],"tasks.":[25,239],"We":[26,88],"present":[27],"IHUBERT,":[28],"a":[29,43,64,91],"monolingual":[30],"PLM":[32],"trained":[33],"from":[34],"scratch":[35],"with":[36],"RoBERTa-base":[38],"encoder":[39],"(125M":[40],"parameters)":[41],"on":[42,95,111,141,146,160,175,179,191,197],"45":[44],"GB":[45],"curated":[46,228],"subset":[47],"Sepahr-Danesh":[50],"collection":[51],"(about":[52],"7-8B":[53],"tokens).":[54],"To":[55],"improve":[56],"corpus":[57,99,201],"quality":[58],"reduce":[60],"redundancy,":[61],"we":[62],"employ":[63],"multi-stage":[65],"preprocessing":[66],"pipeline":[67],"that":[68,203],"includes":[69],"normalization,":[70],"exact":[71],"near-duplicate":[73],"removal,":[74],"anonymization,":[75],"vector-database-based":[77],"semantic":[78],"deduplication":[79],"for":[80],"distribution":[81],"balancing":[82],"control":[83],"across":[84,234],"domains":[85],"registers.":[87],"additionally":[89],"train":[90],"139k-vocabulary":[92],"BPE":[93,204],"tokenizer":[94,195],"full":[97],"to":[100],"better":[101],"capture":[102],"morphology":[104],"orthographic":[106],"variation.":[107],"IHUBERT":[108,136,199,221],"is":[109],"evaluated":[110],"seven":[112],"NLU":[114],"benchmarks":[115],"covering":[116],"NER,":[117],"sentiment":[118],"analysis,":[119],"topic":[120,167],"classification,":[121,168],"NLI,":[122],"extractive":[123,142],"question":[124],"answering,":[125],"relation":[127,182],"extraction,":[128],"using":[129],"task-standard":[130],"metrics":[131],"(entity-level":[132],"F1,":[133],"Macro-F1,":[134],"EM/F1).":[135],"achieves":[137],"its":[138],"strongest":[139],"gains":[140],"QA,":[143],"ranking":[144],"first":[145],"both":[147,235],"PQuAD":[148],"(F1":[149,153],"88.3542)":[150],"ParsiNLU-RC":[152],"49.0987),":[154],"attains":[156],"best":[158],"result":[159],"FarsTail":[161],"(Macro-F1":[162],"0.8350).":[163],"On":[164],"it":[169],"remains":[170,184],"competitive":[171],"(e.g.,":[172],"0.8308":[173],"F1":[174],"ParsTwiNER;":[176],"0.7953":[177],"Macro-F1":[178,190],"DigiMag),":[180],"while":[181],"extraction":[183],"main":[186],"remaining":[187],"gap":[188],"(0.6684":[189],"PERLEX).":[192],"A":[193],"controlled":[194],"ablation":[196],"shows":[202],"yields":[205],"slightly":[206],"lower":[207],"subword":[208],"fragmentation":[209],"than":[210],"WordPiece":[211],"at":[212],"matched":[213],"vocabulary":[214],"size,":[215],"supporting":[216],"our":[217],"tokenization":[218],"design.":[219],"Overall,":[220],"advances":[222],"modeling":[225],"through":[226],"semantically":[227],"large-scale":[229],"broad":[232],"comprehension-oriented":[238]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-06-20T00:00:00"}