{"id":"https://openalex.org/W7116391219","doi":"https://doi.org/10.48550/arxiv.2512.16883","title":"AdaSearch: Balancing Parametric Knowledge and Search in Large Language Models via Reinforcement Learning","display_name":"AdaSearch: Balancing Parametric Knowledge and Search in Large Language Models via Reinforcement Learning","publication_year":2025,"publication_date":"2025-12-18","ids":{"openalex":"https://openalex.org/W7116391219","doi":"https://doi.org/10.48550/arxiv.2512.16883"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2512.16883","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2512.16883","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.2512.16883","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5078048189","display_name":"Tzu-Han Lin","orcid":"https://orcid.org/0000-0002-0849-299X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lin, Tzu-Han","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5120900174","display_name":"Wei-Lin Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Wei-Lin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5120898687","display_name":"Chen-An Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Chen-An","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5120943602","display_name":"Hung-yi Lee","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lee, Hung-yi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5120932270","display_name":"Yun-Nung Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Yun-Nung","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5120871820","display_name":"Yu Meng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Meng, Yu","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":0.20229999721050262,"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":0.20229999721050262,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.12939999997615814,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/T13274","display_name":"Expert finding and Q&A systems","score":0.07919999957084656,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/reinforcement-learning","display_name":"Reinforcement learning","score":0.7261000275611877},{"id":"https://openalex.org/keywords/parametric-statistics","display_name":"Parametric statistics","score":0.513700008392334},{"id":"https://openalex.org/keywords/process","display_name":"Process (computing)","score":0.505299985408783},{"id":"https://openalex.org/keywords/task","display_name":"Task (project management)","score":0.4424000084400177},{"id":"https://openalex.org/keywords/transparency","display_name":"Transparency (behavior)","score":0.38260000944137573},{"id":"https://openalex.org/keywords/decision-process","display_name":"Decision process","score":0.3725000023841858},{"id":"https://openalex.org/keywords/simple","display_name":"Simple (philosophy)","score":0.3149999976158142}],"concepts":[{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.7261000275611877},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6992999911308289},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5648999810218811},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.526199996471405},{"id":"https://openalex.org/C117251300","wikidata":"https://www.wikidata.org/wiki/Q1849855","display_name":"Parametric statistics","level":2,"score":0.513700008392334},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.505299985408783},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.4424000084400177},{"id":"https://openalex.org/C2780233690","wikidata":"https://www.wikidata.org/wiki/Q535347","display_name":"Transparency (behavior)","level":2,"score":0.38260000944137573},{"id":"https://openalex.org/C2984634286","wikidata":"https://www.wikidata.org/wiki/Q1331926","display_name":"Decision process","level":2,"score":0.3725000023841858},{"id":"https://openalex.org/C2780586882","wikidata":"https://www.wikidata.org/wiki/Q7520643","display_name":"Simple (philosophy)","level":2,"score":0.3149999976158142},{"id":"https://openalex.org/C106189395","wikidata":"https://www.wikidata.org/wiki/Q176789","display_name":"Markov decision process","level":3,"score":0.310699999332428},{"id":"https://openalex.org/C28901747","wikidata":"https://www.wikidata.org/wiki/Q177571","display_name":"Decision theory","level":2,"score":0.2915000021457672},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.27379998564720154},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.27219998836517334},{"id":"https://openalex.org/C201789804","wikidata":"https://www.wikidata.org/wiki/Q2362762","display_name":"Search problem","level":2,"score":0.2667999863624573},{"id":"https://openalex.org/C21782646","wikidata":"https://www.wikidata.org/wiki/Q841666","display_name":"Search cost","level":2,"score":0.2662999927997589},{"id":"https://openalex.org/C107327155","wikidata":"https://www.wikidata.org/wiki/Q330268","display_name":"Decision support system","level":2,"score":0.26030001044273376}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2512.16883","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2512.16883","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.2512.16883","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2512.16883","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":[{"id":"https://metadata.un.org/sdg/16","display_name":"Peace, Justice and strong institutions","score":0.7766696810722351}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Equipping":[0],"large":[1],"language":[2],"models":[3],"(LLMs)":[4],"with":[5,57],"search":[6,20,25,61,68,131,220],"engines":[7],"via":[8,133],"reinforcement":[9],"learning":[10],"(RL)":[11],"has":[12],"emerged":[13],"as":[14,142,191],"an":[15,134],"effective":[16],"approach":[17],"for":[18,187],"building":[19],"agents.":[21],"However,":[22,79],"overreliance":[23],"on":[24,40],"introduces":[26],"unnecessary":[27,110,219],"cost":[28],"and":[29,90,109,175,181,193,209,226],"risks":[30,43],"exposure":[31],"to":[32,49,172],"noisy":[33],"or":[34],"malicious":[35],"content,":[36],"while":[37],"relying":[38],"solely":[39,103],"parametric":[41,55,148],"knowledge":[42,56],"hallucination.":[44],"The":[45],"central":[46],"challenge":[47],"is":[48,185,198],"develop":[50],"agents":[51,95,132],"that":[52,96,139,163,212],"adaptively":[53],"balance":[54],"external":[58],"search,":[59,111,174],"invoking":[60],"only":[62],"when":[63],"necessary.":[64],"Prior":[65],"work":[66],"mitigates":[67],"overuse":[69],"by":[70,94,151,201],"shaping":[71],"rewards":[72],"around":[73],"the":[74,113,126,168],"number":[75],"of":[76,115,129,170],"tool":[77],"calls.":[78,99],"these":[80,121,152],"penalties":[81],"require":[82],"substantial":[83],"reward":[84],"engineering,":[85],"provide":[86],"ambiguous":[87],"credit":[88],"assignment,":[89],"can":[91],"be":[92],"exploited":[93],"superficially":[97],"reduce":[98],"Moreover,":[100],"evaluating":[101],"performance":[102],"through":[104],"call":[105],"counts":[106],"conflates":[107],"necessary":[108],"obscuring":[112],"measurement":[114],"true":[116],"adaptive":[117],"behavior.":[118],"To":[119],"address":[120],"limitations,":[122],"we":[123,154],"first":[124],"quantify":[125],"self-knowledge":[127],"awareness":[128],"existing":[130],"F1-based":[135],"decision":[136,169,178,231],"metric,":[137],"revealing":[138],"methods":[140],"such":[141,190],"Search-R1":[143],"often":[144],"overlook":[145],"readily":[146],"available":[147],"knowledge.":[149],"Motivated":[150],"findings,":[153],"propose":[155],"AdaSearch,":[156],"a":[157],"simple":[158],"two-stage,":[159],"outcome-driven":[160],"RL":[161],"framework":[162],"disentangles":[164],"problem":[165],"solving":[166],"from":[167],"whether":[171],"invoke":[173],"makes":[176],"this":[177],"process":[179],"explicit":[180],"interpretable.":[182],"This":[183],"transparency":[184],"crucial":[186],"high-stakes":[188],"domains":[189],"finance":[192],"medical":[194],"question":[195],"answering,":[196],"yet":[197],"largely":[199],"neglected":[200],"prior":[202],"approaches.":[203],"Experiments":[204],"across":[205],"multiple":[206],"model":[207],"families":[208],"sizes":[210],"demonstrate":[211],"AdaSearch":[213],"substantially":[214],"improves":[215],"knowledge-boundary":[216],"awareness,":[217],"reduces":[218],"calls,":[221],"preserves":[222],"strong":[223],"task":[224],"performance,":[225],"offers":[227],"more":[228],"transparent,":[229],"interpretable":[230],"behaviors.":[232]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2025-12-21T00:00:00"}