{"id":"https://openalex.org/W7159058548","doi":"https://doi.org/10.48550/arxiv.2604.26170","title":"EvoSelect: Data-Efficient LLM Evolution for Targeted Task Adaptation","display_name":"EvoSelect: Data-Efficient LLM Evolution for Targeted Task Adaptation","publication_year":2026,"publication_date":"2026-04-28","ids":{"openalex":"https://openalex.org/W7159058548","doi":"https://doi.org/10.48550/arxiv.2604.26170"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2604.26170","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.26170","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":"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.2604.26170","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5134878660","display_name":"Ting-Wei Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Ting-Wei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134924826","display_name":"Sirui Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Sirui","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134899775","display_name":"Jiaru Zou","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zou, Jiaru","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5114058811","display_name":"Yingbing Huang","orcid":"https://orcid.org/0009-0004-8244-2917"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Huang, Yingbing","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134907423","display_name":"Tianxin Wei","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wei, Tianxin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134890686","display_name":"Jingrui He","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"He, Jingrui","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5134884233","display_name":"Hanghang Tong","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tong, Hanghang","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":0.25839999318122864,"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":0.25839999318122864,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.23019999265670776,"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/T11636","display_name":"Artificial Intelligence in Healthcare and Education","score":0.13369999825954437,"subfield":{"id":"https://openalex.org/subfields/2718","display_name":"Health Informatics"},"field":{"id":"https://openalex.org/fields/27","display_name":"Medicine"},"domain":{"id":"https://openalex.org/domains/4","display_name":"Health Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/task","display_name":"Task (project management)","score":0.593500018119812},{"id":"https://openalex.org/keywords/process","display_name":"Process (computing)","score":0.5608000159263611},{"id":"https://openalex.org/keywords/scalability","display_name":"Scalability","score":0.541100025177002},{"id":"https://openalex.org/keywords/adaptation","display_name":"Adaptation (eye)","score":0.5333999991416931},{"id":"https://openalex.org/keywords/iterative-and-incremental-development","display_name":"Iterative and incremental development","score":0.44589999318122864},{"id":"https://openalex.org/keywords/interleaving","display_name":"Interleaving","score":0.4343999922275543},{"id":"https://openalex.org/keywords/synthetic-data","display_name":"Synthetic data","score":0.41359999775886536},{"id":"https://openalex.org/keywords/selection","display_name":"Selection (genetic algorithm)","score":0.4027999937534332}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7824000120162964},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.593500018119812},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.5608000159263611},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.541100025177002},{"id":"https://openalex.org/C139807058","wikidata":"https://www.wikidata.org/wiki/Q352374","display_name":"Adaptation (eye)","level":2,"score":0.5333999991416931},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4982999861240387},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4794999957084656},{"id":"https://openalex.org/C143587482","wikidata":"https://www.wikidata.org/wiki/Q1543216","display_name":"Iterative and incremental development","level":2,"score":0.44589999318122864},{"id":"https://openalex.org/C28034677","wikidata":"https://www.wikidata.org/wiki/Q17092530","display_name":"Interleaving","level":2,"score":0.4343999922275543},{"id":"https://openalex.org/C160920958","wikidata":"https://www.wikidata.org/wiki/Q7662746","display_name":"Synthetic data","level":2,"score":0.41359999775886536},{"id":"https://openalex.org/C81917197","wikidata":"https://www.wikidata.org/wiki/Q628760","display_name":"Selection (genetic algorithm)","level":2,"score":0.4027999937534332},{"id":"https://openalex.org/C158154518","wikidata":"https://www.wikidata.org/wiki/Q7310970","display_name":"Relevance (law)","level":2,"score":0.3634999990463257},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.3628999888896942},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.3628000020980835},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.36169999837875366},{"id":"https://openalex.org/C93959086","wikidata":"https://www.wikidata.org/wiki/Q6888345","display_name":"Model selection","level":2,"score":0.32280001044273376},{"id":"https://openalex.org/C175154964","wikidata":"https://www.wikidata.org/wiki/Q380077","display_name":"Task analysis","level":3,"score":0.32280001044273376},{"id":"https://openalex.org/C2780440489","wikidata":"https://www.wikidata.org/wiki/Q5227278","display_name":"Data-driven","level":2,"score":0.296099990606308},{"id":"https://openalex.org/C97931131","wikidata":"https://www.wikidata.org/wiki/Q5282087","display_name":"Discriminative model","level":2,"score":0.2800000011920929},{"id":"https://openalex.org/C2776436953","wikidata":"https://www.wikidata.org/wiki/Q5163215","display_name":"Consistency (knowledge bases)","level":2,"score":0.26159998774528503},{"id":"https://openalex.org/C153083717","wikidata":"https://www.wikidata.org/wiki/Q6535263","display_name":"Leverage (statistics)","level":2,"score":0.2583000063896179},{"id":"https://openalex.org/C28006648","wikidata":"https://www.wikidata.org/wiki/Q6934509","display_name":"Multi-task learning","level":3,"score":0.25060001015663147}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2604.26170","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.26170","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":"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.2604.26170","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.26170","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":"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":[{"id":"https://metadata.un.org/sdg/10","score":0.5367093086242676,"display_name":"Reduced inequalities"},{"id":"https://metadata.un.org/sdg/16","score":0.401197224855423,"display_name":"Peace, Justice and strong institutions"}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Adapting":[0],"large":[1],"language":[2],"models":[3],"(LLMs)":[4],"to":[5,33,41,139,152],"a":[6,13,25,44,52,125,135,149,204],"targeted":[7,26,102,171,196,225],"task":[8,103,172,178,197],"efficiently":[9],"and":[10,39,54,82,115,174,217],"effectively":[11],"remains":[12],"fundamental":[14],"challenge.":[15],"Such":[16],"adaptation":[17,244],"often":[18],"requires":[19],"iteratively":[20],"improving":[21],"the":[22,75,83,101,161,195],"model":[23,76,118,140],"toward":[24,224],"task,":[27],"yet":[28],"collecting":[29],"high-quality":[30],"human-labeled":[31],"data":[32,47,68,81,109,162,167,239,248],"support":[34],"this":[35,144],"process":[36,84],"is":[37,60,77,85],"costly":[38],"difficult":[40],"scale.":[42],"As":[43],"result,":[45],"synthetic":[46],"generation":[48],"has":[49],"emerged":[50],"as":[51],"flexible":[53],"scalable":[55],"alternative.":[56],"One":[57],"straightforward":[58],"approach":[59],"through":[61,71,180],"an":[62,72,129],"iterative":[63,130],"generation-training":[64],"loop,":[65,132],"where":[66],"candidate":[67,157,191],"are":[69],"synthesized":[70],"external":[73],"generator,":[74,163],"updated":[78],"using":[79],"these":[80],"repeated":[86],"over":[87,246],"iterations.":[88],"However,":[89],"generated":[90],"samples":[91,158,192],"can":[92,110],"be":[93],"noisy,":[94],"highly":[95],"redundant,":[96],"or":[97,237],"even":[98,116],"misaligned":[99],"with":[100,183,194,234],"distribution.":[104,198],"Training":[105],"indiscriminately":[106],"on":[107,143,229],"such":[108],"dilute":[111],"useful":[112],"learning":[113],"signals":[114],"degrade":[117],"performance.":[119],"To":[120,199],"address":[121],"this,":[122],"we":[123,146,202],"introduce":[124],"refined":[126],"paradigm,":[127,145],"namely":[128],"generation-selection-training":[131],"which":[133,187],"incorporates":[134],"selection":[136,249],"step":[137],"prior":[138],"updates.":[141],"Building":[142],"propose":[147],"EvoSelect,":[148],"data-efficient":[150],"framework":[151],"evolve":[153],"LLM":[154,222],"effectively.":[155],"Given":[156],"produced":[159],"by":[160,168],"EvoSelect":[164,219,241],"selects":[165],"training":[166,212],"jointly":[169],"modeling":[170],"alignment":[173,216],"diversity.":[175],"We":[176],"estimate":[177],"relevance":[179],"optimal":[181],"transport":[182],"proxy":[184],"gradient":[185],"representations,":[186],"quantifies":[188],"how":[189],"well":[190],"align":[193],"mitigate":[200],"redundancy,":[201],"incorporate":[203],"diversification":[205],"mechanism":[206],"that":[207,233],"promotes":[208],"coverage":[209],"of":[210],"complementary":[211],"samples.":[213],"By":[214],"interleaving":[215],"diversification,":[218],"enables":[220],"progressive":[221],"evolution":[223],"tasks.":[226],"Extensive":[227],"experiments":[228],"various":[230],"benchmarks":[231],"demonstrate":[232],"either":[235],"weak":[236],"strong":[238],"generators,":[240],"consistently":[242],"improves":[243],"efficacy":[245],"existing":[247],"methods.":[250]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2026-05-01T00:00:00"}