{"id":"https://openalex.org/W4415328522","doi":"https://doi.org/10.48550/arxiv.2505.15117","title":"An Empirical Study on Reinforcement Learning for Reasoning-Search Interleaved LLM Agents","display_name":"An Empirical Study on Reinforcement Learning for Reasoning-Search Interleaved LLM Agents","publication_year":2025,"publication_date":"2025-05-21","ids":{"openalex":"https://openalex.org/W4415328522","doi":"https://doi.org/10.48550/arxiv.2505.15117"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2505.15117","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2505.15117","pdf_url":"https://arxiv.org/pdf/2505.15117","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,"listed_in":[],"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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2505.15117","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5110948473","display_name":"Bowen Jin","orcid":"https://orcid.org/0000-0002-9958-5824","observed_orcids":["https://orcid.org/0000-0002-9958-5824","https://orcid.org/0000-0003-1295-2829"]},"institutions":[{"id":"https://openalex.org/I157725225","display_name":"University of Illinois Urbana-Champaign","ror":"https://ror.org/047426m28","country_code":"US","type":"education","lineage":["https://openalex.org/I157725225"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Jin, Bowen","raw_affiliation_strings":["University of Illinois at Urbana-Champaign"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Illinois at Urbana-Champaign","institution_ids":["https://openalex.org/I157725225"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5002289527","display_name":"Jinsung Yoon","orcid":"https://orcid.org/0000-0002-5481-5171","observed_orcids":["https://orcid.org/0000-0002-5481-5171","https://orcid.org/0009-0002-0954-3637"]},"institutions":[{"id":"https://openalex.org/I1291425158","display_name":"Google (United States)","ror":"https://ror.org/00njsd438","country_code":"US","type":"company","lineage":["https://openalex.org/I1291425158","https://openalex.org/I4210128969"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Yoon, Jinsung","raw_affiliation_strings":["Google Cloud AI Research"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Google Cloud AI Research","institution_ids":["https://openalex.org/I1291425158"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5023566731","display_name":"Priyanka Kargupta","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I157725225","display_name":"University of Illinois Urbana-Champaign","ror":"https://ror.org/047426m28","country_code":"US","type":"education","lineage":["https://openalex.org/I157725225"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Kargupta, Priyanka","raw_affiliation_strings":["University of Illinois at Urbana-Champaign"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Illinois at Urbana-Champaign","institution_ids":["https://openalex.org/I157725225"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5037420650","display_name":"Sercan \u00d6. Ar\u0131k","orcid":"https://orcid.org/0000-0001-6333-1729","observed_orcids":["https://orcid.org/0000-0001-6333-1729"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Arik, Sercan O.","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5019539533","display_name":"Jiawei Han","orcid":"https://orcid.org/0000-0002-3629-2696","observed_orcids":["https://orcid.org/0000-0002-3629-2696"]},"institutions":[{"id":"https://openalex.org/I157725225","display_name":"University of Illinois Urbana-Champaign","ror":"https://ror.org/047426m28","country_code":"US","type":"education","lineage":["https://openalex.org/I157725225"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Han, Jiawei","raw_affiliation_strings":["University of Illinois at Urbana-Champaign"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Illinois at Urbana-Champaign","institution_ids":["https://openalex.org/I157725225"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"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/T10286","display_name":"Information Retrieval and Search Behavior","score":0.9688000082969666,"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":0.9688000082969666,"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/T10028","display_name":"Topic Modeling","score":0.012199999764561653,"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/T10462","display_name":"Reinforcement Learning in Robotics","score":0.004399999976158142,"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/reward-function-design","display_name":"reward function design","score":0.7910000085830688},{"id":"https://openalex.org/keywords/agent-robustness","display_name":"agent robustness","score":0.7429999709129333}],"concepts":[{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.7605999708175659},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7175999879837036},{"id":"https://openalex.org/C114466953","wikidata":"https://www.wikidata.org/wiki/Q6034165","display_name":"Initialization","level":2,"score":0.673799991607666},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5442000031471252},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5360000133514404},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.5156000256538391},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.5156000256538391},{"id":"https://openalex.org/C120936955","wikidata":"https://www.wikidata.org/wiki/Q2155640","display_name":"Empirical research","level":2,"score":0.4754999876022339},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.41519999504089355},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.31369999051094055},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.2881999909877777},{"id":"https://openalex.org/C13687954","wikidata":"https://www.wikidata.org/wiki/Q4826847","display_name":"Autonomous agent","level":2,"score":0.27399998903274536},{"id":"https://openalex.org/C166052673","wikidata":"https://www.wikidata.org/wiki/Q83021","display_name":"Empirical evidence","level":2,"score":0.25929999351501465},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.25119999051094055}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2505.15117","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2505.15117","pdf_url":"https://arxiv.org/pdf/2505.15117","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,"listed_in":[],"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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2505.15117","doi":"https://doi.org/10.48550/arxiv.2505.15117","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2505.15117","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,"listed_in":[],"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"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2505.15117","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2505.15117","pdf_url":"https://arxiv.org/pdf/2505.15117","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,"listed_in":[],"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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[{"id":"https://openalex.org/G2063335192","display_name":"Molecule Maker Lab Institute (MMLI): An AI Institute for Molecular Discovery, Synthetic Strategy, and Manufacturing","funder_award_id":"2019897","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"},{"id":"https://openalex.org/G3178773469","display_name":"III: Medium: Collaborative Research: Mining and Leveraging Knowledge Hypercubes for Complex Applications","funder_award_id":"1956151","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"},{"id":"https://openalex.org/G3269309973","display_name":"HDR Institute: Geospatial Understanding through an Integrative Discovery Environment","funder_award_id":"2118329","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"},{"id":"https://openalex.org/G3904442369","display_name":null,"funder_award_id":"HR0011-24-3-0325","funder_id":"https://openalex.org/F4320332180","funder_display_name":"Defense Advanced Research Projects Agency"},{"id":"https://openalex.org/G6256039072","display_name":"Collaborative Research: III: Medium: Retrieval-Enhanced Machine Learning Through an Information Retrieval Lens","funder_award_id":"2402873","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"}],"funders":[{"id":"https://openalex.org/F4320306076","display_name":"National Science Foundation","ror":"https://ror.org/021nxhr62"},{"id":"https://openalex.org/F4320307791","display_name":"Cisco Systems","ror":"https://ror.org/03yt1ez60"},{"id":"https://openalex.org/F4320332180","display_name":"Defense Advanced Research Projects Agency","ror":"https://ror.org/02caytj08"},{"id":"https://openalex.org/F4320337345","display_name":"Office of Naval Research","ror":"https://ror.org/00rk2pe57"}],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4415328522.pdf","grobid_xml":"https://content.openalex.org/works/W4415328522.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W7153140238","https://openalex.org/W4417422257","https://openalex.org/W4416256501","https://openalex.org/W4416075279","https://openalex.org/W7093085042","https://openalex.org/W6929287018","https://openalex.org/W6910528035","https://openalex.org/W4415315432","https://openalex.org/W6891861597","https://openalex.org/W7139916641"],"abstract_inverted_index":{"Reinforcement":[0],"learning":[1],"(RL)":[2],"has":[3,24],"demonstrated":[4],"strong":[5],"potential":[6],"in":[7,89,122,157,183],"training":[8,47,160],"large":[9],"language":[10],"models":[11],"(LLMs)":[12],"capable":[13],"of":[14,44,55,77,85,137,150,165],"complex":[15],"reasoning":[16,36],"for":[17,46,175],"real-world":[18,184],"problem":[19],"solving.":[20],"More":[21],"recently,":[22],"RL":[23,45,91,145,159],"been":[25],"leveraged":[26],"to":[27,105],"create":[28],"sophisticated":[29],"LLM-based":[30,180],"search":[31,38,48,87,151,181],"agents":[32,49,57,182],"that":[33],"adeptly":[34],"combine":[35],"with":[37],"engine":[39,88,152],"use.":[40],"While":[41],"the":[42,52,73,78,83,86,90,133,138,148,163,166],"use":[43],"is":[50,187],"promising,":[51],"optimal":[53],"design":[54],"such":[56,67],"remains":[58],"not":[59],"fully":[60],"understood.":[61],"In":[62,97],"particular,":[63],"key":[64,116],"factors":[65],"--":[66,93],"as":[68],"(1)":[69],"reward":[70],"formulation,":[71],"(2)":[72],"choice":[74,149],"and":[75,81,109,135,147,162,178],"characteristics":[76],"underlying":[79],"LLM,":[80],"(3)":[82],"role":[84,156],"process":[92],"require":[94],"further":[95],"investigation.":[96],"this":[98],"work,":[99],"we":[100],"conduct":[101],"comprehensive":[102],"empirical":[103],"studies":[104],"systematically":[106],"investigate":[107],"these":[108],"offer":[110],"actionable":[111],"insights.":[112],"We":[113],"highlight":[114],"several":[115],"findings:":[117],"format":[118],"rewards":[119,129],"are":[120],"effective":[121],"improving":[123],"final":[124],"performance,":[125],"whereas":[126],"intermediate":[127],"retrieval":[128],"have":[130],"limited":[131],"impact;":[132],"scale":[134],"initialization":[136],"LLM":[139],"(general-purpose":[140],"vs.":[141],"reasoning-specialized)":[142],"significantly":[143],"influence":[144],"outcomes;":[146],"plays":[153],"a":[154],"critical":[155],"shaping":[158],"dynamics":[161],"robustness":[164],"trained":[167],"agent":[168],"during":[169],"inference.":[170],"These":[171],"establish":[172],"important":[173],"guidelines":[174],"successfully":[176],"building":[177],"deploying":[179],"applications.":[185],"Code":[186],"available":[188],"at":[189],"https://github.com/PeterGriffinJin/Search-R1.":[190]},"counts_by_year":[],"updated_date":"2026-09-30T09:27:27.532063","created_date":"2025-10-19T00:00:00"}