{"id":"https://openalex.org/W7204601596","doi":"https://doi.org/10.48550/arxiv.2608.26530","title":"PILOT in the Loop: Live Self-Improvement for Long-Horizon Agents","display_name":"PILOT in the Loop: Live Self-Improvement for Long-Horizon Agents","publication_year":2026,"publication_date":"2026-08-27","ids":{"openalex":"https://openalex.org/W7204601596","doi":"https://doi.org/10.48550/arxiv.2608.26530"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2608.26530","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.26530","pdf_url":"https://arxiv.org/pdf/2608.26530","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/2608.26530","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5148522434","display_name":"Yang Xiao","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xiao, Yang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139138548","display_name":"Yusong Sun","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Sun, Yusong","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5148510340","display_name":"Haoyi Wu","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wu, Haoyi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5074113705","display_name":"Wenyang Hui","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hui, Wenyang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5148500951","display_name":"Wen Da","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Da, Wen","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5025908613","display_name":"Zhaokai Luo","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Luo, Zhaokai","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5148448611","display_name":"Mu Chuan","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chuan, Mu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5146859957","display_name":"Yao Hu","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hu, Yao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5148519190","display_name":"Wenjie Li","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Wenjie","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5148535573","display_name":"Chengyue Jiang","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jiang, Chengyue","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":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/T10906","display_name":"AI-based Problem Solving and Planning","score":0.5259000062942505,"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/T10906","display_name":"AI-based Problem Solving and Planning","score":0.5259000062942505,"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/T13904","display_name":"Artificial Intelligence Applications","score":0.07259999960660934,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.0494999997317791,"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/long-horizon-agents","display_name":"long-horizon agents","score":0.9319999814033508},{"id":"https://openalex.org/keywords/agent-harnesses","display_name":"agent harnesses","score":0.531000018119812},{"id":"https://openalex.org/keywords/large-language-models","display_name":"large language models","score":0.5450000166893005}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7257999777793884},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.6621000170707703},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.6585000157356262},{"id":"https://openalex.org/C2779110517","wikidata":"https://www.wikidata.org/wiki/Q1240788","display_name":"Supervisor","level":2,"score":0.6449999809265137},{"id":"https://openalex.org/C13662910","wikidata":"https://www.wikidata.org/wiki/Q193139","display_name":"Trajectory","level":2,"score":0.3959999978542328},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.3537999987602234},{"id":"https://openalex.org/C86532276","wikidata":"https://www.wikidata.org/wiki/Q1184065","display_name":"Delegation","level":2,"score":0.3393999934196472},{"id":"https://openalex.org/C2777549182","wikidata":"https://www.wikidata.org/wiki/Q15117556","display_name":"Abort","level":2,"score":0.3249000012874603},{"id":"https://openalex.org/C2778476105","wikidata":"https://www.wikidata.org/wiki/Q628539","display_name":"Workload","level":2,"score":0.31029999256134033},{"id":"https://openalex.org/C142944206","wikidata":"https://www.wikidata.org/wiki/Q1786137","display_name":"Proactivity","level":2,"score":0.30709999799728394},{"id":"https://openalex.org/C115903868","wikidata":"https://www.wikidata.org/wiki/Q80993","display_name":"Software engineering","level":1,"score":0.30169999599456787},{"id":"https://openalex.org/C2778744189","wikidata":"https://www.wikidata.org/wiki/Q4677551","display_name":"Active database","level":2,"score":0.29980000853538513},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.296999990940094},{"id":"https://openalex.org/C107457646","wikidata":"https://www.wikidata.org/wiki/Q207434","display_name":"Human\u2013computer interaction","level":1,"score":0.29260000586509705},{"id":"https://openalex.org/C175154964","wikidata":"https://www.wikidata.org/wiki/Q380077","display_name":"Task analysis","level":3,"score":0.29010000824928284},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.2881999909877777},{"id":"https://openalex.org/C2989134064","wikidata":"https://www.wikidata.org/wiki/Q288510","display_name":"Execution time","level":2,"score":0.2865999937057495},{"id":"https://openalex.org/C148220186","wikidata":"https://www.wikidata.org/wiki/Q7111912","display_name":"Outcome (game theory)","level":2,"score":0.2816999852657318},{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","level":1,"score":0.26600000262260437},{"id":"https://openalex.org/C165136773","wikidata":"https://www.wikidata.org/wiki/Q1363179","display_name":"Single point of failure","level":2,"score":0.26339998841285706},{"id":"https://openalex.org/C44154836","wikidata":"https://www.wikidata.org/wiki/Q45045","display_name":"Simulation","level":1,"score":0.2508000135421753}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2608.26530","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.26530","pdf_url":"https://arxiv.org/pdf/2608.26530","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.2608.26530","doi":"https://doi.org/10.48550/arxiv.2608.26530","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2608.26530","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:2608.26530","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.26530","pdf_url":"https://arxiv.org/pdf/2608.26530","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":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7204601596.pdf","grobid_xml":"https://content.openalex.org/works/W7204601596.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W7163683658","https://openalex.org/W7215035621","https://openalex.org/W7221796217","https://openalex.org/W7203696678","https://openalex.org/W7220938708","https://openalex.org/W7160900680","https://openalex.org/W4415949658","https://openalex.org/W7206206100","https://openalex.org/W7214762678","https://openalex.org/W4416400958"],"abstract_inverted_index":{"Long-horizon":[0],"agent":[1,65],"runs":[2],"generate":[3],"experience":[4,20,51],"that":[5,43],"can":[6],"improve":[7],"both":[8,52],"the":[9,29,55,61,119,170],"current":[10],"run":[11,31,57],"and":[12,35,58,78,124,130,139,145,179,190,202],"future":[13],"work.":[14],"Most":[15],"self-improvement":[16,44,104,171],"methods":[17],"process":[18],"this":[19,71],"only":[21],"after":[22],"execution":[23,77,88,135],"ends,":[24],"so":[25],"they":[26],"cannot":[27,91],"redirect":[28,54,92,116],"active":[30,56,94,120],"or":[32,117],"immediately":[33],"apply":[34],"validate":[36],"lessons":[37],"learned":[38],"from":[39],"it.":[40],"We":[41,96],"argue":[42],"should":[45],"instead":[46],"be":[47],"live,":[48],"using":[49],"emerging":[50],"to":[53,59,165],"update":[60],"persistent":[62],"harness.":[63],"Existing":[64],"architectures":[66],"do":[67],"not":[68],"fully":[69],"support":[70],"goal.":[72],"Single-agent":[73],"self-correction":[74],"combines":[75],"task":[76],"trajectory":[79],"assessment":[80],"within":[81],"one":[82],"context,":[83],"while":[84,192],"subagent":[85],"delegation":[86],"separates":[87],"but":[89],"typically":[90],"an":[93],"subagent.":[95],"present":[97],"PILOT,":[98],"a":[99,113],"supervisor-worker":[100],"harness":[101],"for":[102],"live":[103,110,126],"through":[105],"two":[106,142],"coupled":[107],"mechanisms:":[108],"(1)":[109],"steering":[111],"lets":[112],"separate":[114],"supervisor":[115],"abort":[118],"worker":[121],"during":[122,134],"execution;":[123],"(2)":[125],"self-evolution":[127],"distils":[128],"procedures":[129],"failure":[131],"modes":[132],"revealed":[133],"into":[136],"reusable":[137],"skills":[138],"memory.":[140],"Across":[141],"frozen":[143],"backbones":[144],"three":[146],"benchmarks,":[147],"PILOT":[148,159,173],"ranks":[149],"first":[150],"in":[151],"five":[152],"of":[153],"six":[154],"configurations.":[155],"On":[156],"Terminal-Bench":[157],"2.0,":[158],"outperforms":[160],"counterpart":[161],"harnesses":[162],"by":[163,188,200],"up":[164],"9.8":[166],"percentage":[167],"points.":[168],"In":[169],"setting,":[172],"gains":[174],"14.6":[175],"points":[176,181],"with":[177,182],"GLM-5.1":[178],"12.4":[180],"Kimi-K2.6.":[183],"Mean":[184],"output":[185,197],"tokens":[186,198],"fall":[187],"42.9%":[189],"47.4%,":[191],"successful":[193],"evaluations":[194],"per":[195],"million":[196],"rise":[199],"110.3%":[201],"134.0%,":[203],"respectively.":[204]},"counts_by_year":[],"updated_date":"2026-10-28T10:33:24.094632","created_date":"2026-08-29T00:00:00"}