{"id":"https://openalex.org/W7164554678","doi":"https://doi.org/10.48550/arxiv.2606.12908","title":"SENTINEL: Failure-Driven Reinforcement Learning for Training Tool-Using Language Model Agents","display_name":"SENTINEL: Failure-Driven Reinforcement Learning for Training Tool-Using Language Model Agents","publication_year":2026,"publication_date":"2026-06-11","ids":{"openalex":"https://openalex.org/W7164554678","doi":"https://doi.org/10.48550/arxiv.2606.12908"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2606.12908","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.12908","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.2606.12908","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5138533500","display_name":"Ziyi Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Ziyi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138512971","display_name":"Yuxuan Lu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lu, Yuxuan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138493262","display_name":"Yimeng Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Yimeng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138520670","display_name":"Qun Liu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Qun","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138498596","display_name":"Chen Luo","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Luo, Chen","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138506178","display_name":"Jiri Gesi","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Gesi, Jiri","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5029653356","display_name":"H\u00fcseyin Ozan \u04aairkino\u011flu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lu, Hanqing","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5078860840","display_name":"Yisi Sang","orcid":"https://orcid.org/0000-0002-8876-7542"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Sang, Yisi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138503345","display_name":"Manling Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Manling","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5006022044","display_name":"J W Huang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Huang, Jing","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5138551583","display_name":"Dakuo Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Dakuo","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/T10462","display_name":"Reinforcement Learning in Robotics","score":0.3659000098705292,"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/T10462","display_name":"Reinforcement Learning in Robotics","score":0.3659000098705292,"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":0.10859999805688858,"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":0.05689999833703041,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/reinforcement-learning","display_name":"Reinforcement learning","score":0.8679999709129333},{"id":"https://openalex.org/keywords/task","display_name":"Task (project management)","score":0.6547999978065491},{"id":"https://openalex.org/keywords/executable","display_name":"Executable","score":0.5519000291824341},{"id":"https://openalex.org/keywords/scalability","display_name":"Scalability","score":0.5135999917984009},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.42750000953674316},{"id":"https://openalex.org/keywords/training","display_name":"Training (meteorology)","score":0.39660000801086426},{"id":"https://openalex.org/keywords/task-analysis","display_name":"Task analysis","score":0.3864000141620636}],"concepts":[{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.8679999709129333},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7720999717712402},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.6547999978065491},{"id":"https://openalex.org/C160145156","wikidata":"https://www.wikidata.org/wiki/Q778586","display_name":"Executable","level":2,"score":0.5519000291824341},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.5135999917984009},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.49459999799728394},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.42750000953674316},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.39660000801086426},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.388700008392334},{"id":"https://openalex.org/C175154964","wikidata":"https://www.wikidata.org/wiki/Q380077","display_name":"Task analysis","level":3,"score":0.3864000141620636},{"id":"https://openalex.org/C164155591","wikidata":"https://www.wikidata.org/wiki/Q2067766","display_name":"Satisfiability modulo theories","level":2,"score":0.3643999993801117},{"id":"https://openalex.org/C67203356","wikidata":"https://www.wikidata.org/wiki/Q1321905","display_name":"Reinforcement","level":2,"score":0.3456999957561493},{"id":"https://openalex.org/C107457646","wikidata":"https://www.wikidata.org/wiki/Q207434","display_name":"Human\u2013computer interaction","level":1,"score":0.29829999804496765},{"id":"https://openalex.org/C203479927","wikidata":"https://www.wikidata.org/wiki/Q5165939","display_name":"Controller (irrigation)","level":2,"score":0.27730000019073486},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.27059999108314514},{"id":"https://openalex.org/C2778770139","wikidata":"https://www.wikidata.org/wiki/Q1966904","display_name":"Solver","level":2,"score":0.2689000070095062},{"id":"https://openalex.org/C2775924081","wikidata":"https://www.wikidata.org/wiki/Q55608371","display_name":"Control (management)","level":2,"score":0.2581000030040741}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2606.12908","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.12908","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.2606.12908","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.12908","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":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Language":[0],"model":[1,151,167],"agents":[2,18,32],"are":[3,49],"increasingly":[4,58],"effective":[5,155],"in":[6,21],"solving":[7],"realistic":[8],"tasks":[9,48,111,143],"through":[10],"multi-turn":[11],"tool":[12],"use.":[13],"However,":[14],"training":[15,44,90,161],"reliable":[16],"tool-using":[17,165],"remains":[19],"challenging":[20],"practice.":[22],"While":[23],"reinforcement":[24,79],"learning":[25,80],"provides":[26],"an":[27,154],"on-policy":[28],"paradigm":[29],"for":[30,163],"improving":[31,164],"from":[33,133],"their":[34],"own":[35],"environment":[36],"interactions,":[37],"its":[38],"effectiveness":[39],"depends":[40],"heavily":[41],"on":[42,71,121,140],"the":[43,53,61,84,97,107,117,122],"task":[45,54],"distribution.":[46],"When":[47],"fixed":[50],"before":[51],"training,":[52],"distribution":[55],"can":[56],"become":[57],"mismatched":[59],"with":[60,128],"policy's":[62],"evolving":[63],"capabilities,":[64],"causing":[65],"many":[66],"rollouts":[67],"to":[68,135],"be":[69],"spent":[70],"uninformative":[72],"tasks.":[73,91,124],"We":[74],"propose":[75],"SENTINEL,":[76],"a":[77,94],"failure-driven":[78],"framework":[81],"that":[82,112,150],"turns":[83],"Solver's":[85],"rollout":[86],"failures":[87,152],"into":[88],"targeted":[89,123,160],"SENTINEL":[92,130],"follows":[93],"Controller--Proposer--Solver":[95],"loop:":[96],"Controller":[98],"analyzes":[99],"failed":[100],"trajectories":[101],"and":[102,116,137,156],"summarizes":[103],"recurring":[104],"error":[105],"patterns,":[106],"Proposer":[108],"generates":[109],"executable":[110],"stress":[113],"these":[114],"weaknesses,":[115],"Solver":[118],"is":[119],"trained":[120],"On":[125],"Tau2-Bench":[126],"Retail":[127],"Qwen3-4B-Thinking-2507,":[129],"improves":[131],"Pass\\^{}1":[132],"66.4":[134],"74.9":[136],"outperforms":[138],"RL":[139],"general":[141],"synthetic":[142],"across":[144],"Pass\\^{}k":[145],"metrics.":[146],"These":[147],"results":[148],"demonstrate":[149],"provide":[153],"scalable":[157],"source":[158],"of":[159],"signal":[162],"language":[166],"agents.":[168]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-06-13T00:00:00"}