{"id":"https://openalex.org/W7165120137","doi":"https://doi.org/10.48550/arxiv.2606.18650","title":"BLADE: Scalable Bi-level Adaptive Data Selection for LLM Training","display_name":"BLADE: Scalable Bi-level Adaptive Data Selection for LLM Training","publication_year":2026,"publication_date":"2026-06-17","ids":{"openalex":"https://openalex.org/W7165120137","doi":"https://doi.org/10.48550/arxiv.2606.18650"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2606.18650","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.18650","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.2606.18650","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5138949810","display_name":"Jiaxing Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Jiaxing","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137857614","display_name":"Deping Xiang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xiang, Deping","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138960328","display_name":"Jin Xu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xu, Jin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138957861","display_name":"Zirui Liu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Zirui","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138938084","display_name":"Zicheng Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Zicheng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101998250","display_name":"Guoqiang Gong","orcid":"https://orcid.org/0000-0003-2622-9092"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Gong, Guoqiang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5036603233","display_name":"Ji Fang","orcid":"https://orcid.org/0000-0003-3418-535X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Fang, Jun","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138947626","display_name":"Chao Liu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Chao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5054904119","display_name":"Pengzhang Liu","orcid":"https://orcid.org/0000-0002-6031-5245"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Pengzhang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138898264","display_name":"Tongxuan Liu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Tongxuan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138851509","display_name":"Ke Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Ke","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5138887314","display_name":"Qixia Jiang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jiang, Qixia","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/T12535","display_name":"Machine Learning and Data Classification","score":0.18080000579357147,"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/T12535","display_name":"Machine Learning and Data Classification","score":0.18080000579357147,"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/T10028","display_name":"Topic Modeling","score":0.17800000309944153,"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/T12072","display_name":"Machine Learning and Algorithms","score":0.10750000178813934,"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/scalability","display_name":"Scalability","score":0.6299999952316284},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.5177000164985657},{"id":"https://openalex.org/keywords/selection","display_name":"Selection (genetic algorithm)","score":0.48330000042915344},{"id":"https://openalex.org/keywords/heuristic","display_name":"Heuristic","score":0.4553000032901764},{"id":"https://openalex.org/keywords/construct","display_name":"Construct (python library)","score":0.4302999973297119},{"id":"https://openalex.org/keywords/filter","display_name":"Filter (signal processing)","score":0.3783000111579895},{"id":"https://openalex.org/keywords/model-selection","display_name":"Model selection","score":0.37439998984336853}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.766700029373169},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.6299999952316284},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5511999726295471},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.5177000164985657},{"id":"https://openalex.org/C81917197","wikidata":"https://www.wikidata.org/wiki/Q628760","display_name":"Selection (genetic algorithm)","level":2,"score":0.48330000042915344},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.46959999203681946},{"id":"https://openalex.org/C173801870","wikidata":"https://www.wikidata.org/wiki/Q201413","display_name":"Heuristic","level":2,"score":0.4553000032901764},{"id":"https://openalex.org/C2780801425","wikidata":"https://www.wikidata.org/wiki/Q5164392","display_name":"Construct (python library)","level":2,"score":0.4302999973297119},{"id":"https://openalex.org/C106131492","wikidata":"https://www.wikidata.org/wiki/Q3072260","display_name":"Filter (signal processing)","level":2,"score":0.3783000111579895},{"id":"https://openalex.org/C93959086","wikidata":"https://www.wikidata.org/wiki/Q6888345","display_name":"Model selection","level":2,"score":0.37439998984336853},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.36579999327659607},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.34689998626708984},{"id":"https://openalex.org/C2780440489","wikidata":"https://www.wikidata.org/wiki/Q5227278","display_name":"Data-driven","level":2,"score":0.32260000705718994},{"id":"https://openalex.org/C2778755073","wikidata":"https://www.wikidata.org/wiki/Q10858537","display_name":"Scale (ratio)","level":2,"score":0.3172999918460846},{"id":"https://openalex.org/C148483581","wikidata":"https://www.wikidata.org/wiki/Q446488","display_name":"Feature selection","level":2,"score":0.30480000376701355},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.30079999566078186},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.2775999903678894},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.2736000120639801},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.2727999985218048},{"id":"https://openalex.org/C75684735","wikidata":"https://www.wikidata.org/wiki/Q858810","display_name":"Big data","level":2,"score":0.25040000677108765}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2606.18650","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.18650","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.2606.18650","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.18650","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":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"As":[0],"Large":[1],"Language":[2],"Model":[3],"(LLM)":[4],"datasets":[5],"scale":[6],"to":[7,19,120],"trillions":[8],"of":[9],"tokens,":[10],"data":[11,34,93,123,180],"selection":[12,35,181],"has":[13],"emerged":[14],"as":[15,104,165],"a":[16,68,89,105,117,139,166,184],"critical":[17],"frontier":[18],"filter":[20],"out":[21],"uninformative":[22],"noise":[23],"and":[24],"construct":[25],"adaptive":[26],"learning":[27],"trajectories.":[28],"Beyond":[29],"static":[30,69,135],"heuristic":[31],"filtering,":[32],"advanced":[33],"methods":[36,49,61,103],"for":[37,92,187],"LLM":[38,188],"training":[39],"largely":[40],"follow":[41],"two":[42],"paradigms,":[43],"each":[44],"with":[45,75,138,145],"fundamental":[46],"limitations.":[47],"Influence-based":[48],"provide":[50],"principled":[51,118],"bi-level":[52,98],"objectives":[53],"but":[54,65,132],"require":[55],"intractable":[56],"inverse-Hessian":[57,113],"computations,":[58],"while":[59,115],"excess-loss":[60,121,130],"are":[62],"computationally":[63],"efficient":[64,158],"rely":[66],"on":[67],"reference":[70,136],"model":[71,79,137],"that":[72,142,150,175],"becomes":[73],"misaligned":[74],"the":[76,97,134],"evolving":[77],"proxy":[78],"during":[80],"training.":[81,146,189],"We":[82],"propose":[83],"BLADE":[84,95,164,176],"(Bi-Level":[85],"Adaptive":[86],"Data":[87],"sElection),":[88],"Hessian-free":[90],"framework":[91],"selection.":[94,124],"reformulates":[96],"optimization":[99],"problem":[100],"underlying":[101],"influence-based":[102],"penalized":[106,152],"single-level":[107],"objective":[108,127],"via":[109],"Lagrange":[110],"multipliers,":[111],"avoiding":[112],"computation":[114],"revealing":[116],"connection":[119],"based":[122],"The":[125],"resulting":[126],"recovers":[128],"an":[129],"form":[131],"replaces":[133],"dynamic":[140],"one":[141],"stays":[143],"synchronized":[144],"Theoretically,":[147],"we":[148,162],"prove":[149],"this":[151],"formulation":[153],"guarantees":[154],"first-order":[155],"convergence.":[156],"For":[157],"online":[159],"batch":[160],"selection,":[161],"instantiate":[163],"memoryless":[167],"randomized":[168],"block-coordinate":[169],"Frank-Wolfe":[170],"algorithm.":[171],"Extensive":[172],"experiments":[173],"show":[174],"consistently":[177],"outperforms":[178],"state-of-the-art":[179],"baselines,":[182],"providing":[183],"practical":[185],"recipe":[186]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-06-19T00:00:00"}