{"id":"https://openalex.org/W7172537726","doi":"https://doi.org/10.48550/arxiv.2608.02636","title":"Rethinking Self-Evolving Agent Skills: Feedback Dynamics over Multiple Rounds","display_name":"Rethinking Self-Evolving Agent Skills: Feedback Dynamics over Multiple Rounds","publication_year":2026,"publication_date":"2026-07-31","ids":{"openalex":"https://openalex.org/W7172537726","doi":"https://doi.org/10.48550/arxiv.2608.02636"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2608.02636","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.02636","pdf_url":"https://arxiv.org/pdf/2608.02636","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","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/2608.02636","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5125747332","display_name":"Yuxuan Liu","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I200769079","display_name":"Hong Kong University of Science and Technology","ror":"https://ror.org/00q4vv597","country_code":"HK","type":"education","lineage":["https://openalex.org/I200769079"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Liu, Yuxuan","raw_affiliation_strings":["The Hong Kong University of Science and Technology"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Hong Kong University of Science and Technology","institution_ids":["https://openalex.org/I200769079"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137652523","display_name":"Zhaochen Su","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I200769079","display_name":"Hong Kong University of Science and Technology","ror":"https://ror.org/00q4vv597","country_code":"HK","type":"education","lineage":["https://openalex.org/I200769079"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Su, Zhaochen","raw_affiliation_strings":["The Hong Kong University of Science and Technology"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Hong Kong University of Science and Technology","institution_ids":["https://openalex.org/I200769079"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5122008396","display_name":"Yuhao Zhang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I204983213","display_name":"Harbin Institute of Technology","ror":"https://ror.org/01yqg2h08","country_code":"CN","type":"education","lineage":["https://openalex.org/I204983213"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhang, Yuhao","raw_affiliation_strings":["Harbin Institute of Technology , Shenzhen"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Harbin Institute of Technology , Shenzhen","institution_ids":["https://openalex.org/I204983213"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5124050941","display_name":"Jiahe Guo","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I204983213","display_name":"Harbin Institute of Technology","ror":"https://ror.org/01yqg2h08","country_code":"CN","type":"education","lineage":["https://openalex.org/I204983213"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Guo, Jiahe","raw_affiliation_strings":["Harbin Institute of Technology"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Harbin Institute of Technology","institution_ids":["https://openalex.org/I204983213"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5126889694","display_name":"Zhongwei Xie","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I200769079","display_name":"Hong Kong University of Science and Technology","ror":"https://ror.org/00q4vv597","country_code":"HK","type":"education","lineage":["https://openalex.org/I200769079"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Xie, Zhongwei","raw_affiliation_strings":["The Hong Kong University of Science and Technology"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Hong Kong University of Science and Technology","institution_ids":["https://openalex.org/I200769079"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5144449582","display_name":"Huihao Jing","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I200769079","display_name":"Hong Kong University of Science and Technology","ror":"https://ror.org/00q4vv597","country_code":"HK","type":"education","lineage":["https://openalex.org/I200769079"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Jing, Huihao","raw_affiliation_strings":["The Hong Kong University of Science and Technology"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Hong Kong University of Science and Technology","institution_ids":["https://openalex.org/I200769079"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5029627230","display_name":"Lingyun Xie","orcid":"https://orcid.org/0000-0001-8437-5245","observed_orcids":["https://orcid.org/0000-0001-8437-5245"]},"institutions":[{"id":"https://openalex.org/I200769079","display_name":"Hong Kong University of Science and Technology","ror":"https://ror.org/00q4vv597","country_code":"HK","type":"education","lineage":["https://openalex.org/I200769079"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Xie, Lingyun","raw_affiliation_strings":["The Hong Kong University of Science and Technology"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Hong Kong University of Science and Technology","institution_ids":["https://openalex.org/I200769079"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5093047682","display_name":"Qing Zong","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I200769079","display_name":"Hong Kong University of Science and Technology","ror":"https://ror.org/00q4vv597","country_code":"HK","type":"education","lineage":["https://openalex.org/I200769079"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Zong, Qing","raw_affiliation_strings":["The Hong Kong University of Science and Technology"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Hong Kong University of Science and Technology","institution_ids":["https://openalex.org/I200769079"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5095848529","display_name":"Yauwai Yim","orcid":"https://orcid.org/0009-0002-1458-1208","observed_orcids":["https://orcid.org/0009-0002-1458-1208"]},"institutions":[{"id":"https://openalex.org/I200769079","display_name":"Hong Kong University of Science and Technology","ror":"https://ror.org/00q4vv597","country_code":"HK","type":"education","lineage":["https://openalex.org/I200769079"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Yim, Yauwai","raw_affiliation_strings":["The Hong Kong University of Science and Technology"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Hong Kong University of Science and Technology","institution_ids":["https://openalex.org/I200769079"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5144461790","display_name":"Zhixiong Zhang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I183067930","display_name":"Shanghai Jiao Tong University","ror":"https://ror.org/0220qvk04","country_code":"CN","type":"education","lineage":["https://openalex.org/I183067930"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhang, Zhixiong","raw_affiliation_strings":["Shanghai Jiao Tong University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Shanghai Jiao Tong University","institution_ids":["https://openalex.org/I183067930"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100327744","display_name":"Haoran Li","orcid":"https://orcid.org/0000-0003-1656-1278","observed_orcids":["https://orcid.org/0000-0003-1656-1278"]},"institutions":[{"id":"https://openalex.org/I200769079","display_name":"Hong Kong University of Science and Technology","ror":"https://ror.org/00q4vv597","country_code":"HK","type":"education","lineage":["https://openalex.org/I200769079"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Li, Haoran","raw_affiliation_strings":["The Hong Kong University of Science and Technology"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Hong Kong University of Science and Technology","institution_ids":["https://openalex.org/I200769079"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5020880385","display_name":"Yangqiu Song","orcid":"https://orcid.org/0000-0002-7818-6090","observed_orcids":["https://orcid.org/0000-0002-7818-6090"]},"institutions":[{"id":"https://openalex.org/I200769079","display_name":"Hong Kong University of Science and Technology","ror":"https://ror.org/00q4vv597","country_code":"HK","type":"education","lineage":["https://openalex.org/I200769079"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Song, Yangqiu","raw_affiliation_strings":["The Hong Kong University of Science and Technology"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Hong Kong University of Science and Technology","institution_ids":["https://openalex.org/I200769079"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":3,"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/T11975","display_name":"Evolutionary Algorithms and Applications","score":0.82669997215271,"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/T11975","display_name":"Evolutionary Algorithms and Applications","score":0.82669997215271,"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/T10906","display_name":"AI-based Problem Solving and Planning","score":0.06700000166893005,"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/T10286","display_name":"Information Retrieval and Search Behavior","score":0.020899999886751175,"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/skill-evolution","display_name":"skill evolution","score":0.4059999883174896},{"id":"https://openalex.org/keywords/execution-feedback","display_name":"execution feedback","score":0.7329999804496765},{"id":"https://openalex.org/keywords/success-and-failure-feedback","display_name":"success and failure feedback","score":0.7089999914169312},{"id":"https://openalex.org/keywords/test-time-scaling","display_name":"test-time scaling","score":0.47999998927116394},{"id":"https://openalex.org/keywords/simultaneous-sampling","display_name":"synchronous sampling","score":0.5950000286102295},{"id":"https://openalex.org/keywords/robustness","display_name":"robustness","score":0.5630000233650208},{"id":"https://openalex.org/keywords/transfer","display_name":"transfer","score":0.578000009059906},{"id":"https://openalex.org/keywords/llm-based-agents","display_name":"LLM agents","score":0.4869999885559082}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7001000046730042},{"id":"https://openalex.org/C55166926","wikidata":"https://www.wikidata.org/wiki/Q2892946","display_name":"Oracle","level":2,"score":0.6040999889373779},{"id":"https://openalex.org/C189430467","wikidata":"https://www.wikidata.org/wiki/Q7293293","display_name":"Ranking (information retrieval)","level":2,"score":0.5419999957084656},{"id":"https://openalex.org/C81917197","wikidata":"https://www.wikidata.org/wiki/Q628760","display_name":"Selection (genetic algorithm)","level":2,"score":0.4756999909877777},{"id":"https://openalex.org/C180591056","wikidata":"https://www.wikidata.org/wiki/Q654437","display_name":"Executor","level":2,"score":0.41830000281333923},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.3831999897956848},{"id":"https://openalex.org/C140779682","wikidata":"https://www.wikidata.org/wiki/Q210868","display_name":"Sampling (signal processing)","level":3,"score":0.3714999854564667},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3637999892234802},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.3617999851703644},{"id":"https://openalex.org/C176217482","wikidata":"https://www.wikidata.org/wiki/Q860554","display_name":"Metric (unit)","level":2,"score":0.35370001196861267},{"id":"https://openalex.org/C145912823","wikidata":"https://www.wikidata.org/wiki/Q113558","display_name":"Dynamics (music)","level":2,"score":0.33169999718666077},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.31040000915527344},{"id":"https://openalex.org/C2780598303","wikidata":"https://www.wikidata.org/wiki/Q65921492","display_name":"Flexibility (engineering)","level":2,"score":0.28929999470710754},{"id":"https://openalex.org/C2776207758","wikidata":"https://www.wikidata.org/wiki/Q5303302","display_name":"Downstream (manufacturing)","level":2,"score":0.2849000096321106},{"id":"https://openalex.org/C73602740","wikidata":"https://www.wikidata.org/wiki/Q7795822","display_name":"Thompson sampling","level":3,"score":0.28299999237060547},{"id":"https://openalex.org/C43214815","wikidata":"https://www.wikidata.org/wiki/Q7310987","display_name":"Reliability (semiconductor)","level":3,"score":0.28060001134872437},{"id":"https://openalex.org/C107673813","wikidata":"https://www.wikidata.org/wiki/Q812534","display_name":"Bayesian probability","level":2,"score":0.2597000002861023}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2608.02636","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.02636","pdf_url":"https://arxiv.org/pdf/2608.02636","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2608.02636","doi":"https://doi.org/10.48550/arxiv.2608.02636","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2608.02636","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":"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"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2608.02636","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.02636","pdf_url":"https://arxiv.org/pdf/2608.02636","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"sustainable_development_goals_aurora":[{"id":"https://metadata.un.org/sdg/4","display_name":"Quality education","score":0.6670271754264832}],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7172537726.pdf","grobid_xml":"https://content.openalex.org/works/W7172537726.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W7203491292","https://openalex.org/W7222614344","https://openalex.org/W7213286606","https://openalex.org/W7223038946","https://openalex.org/W7201863843","https://openalex.org/W7201834752","https://openalex.org/W7133528385","https://openalex.org/W7164828440","https://openalex.org/W7211998686","https://openalex.org/W7164396653"],"abstract_inverted_index":{"Self-evolving":[0],"skill":[1,13,126,202,217],"systems":[2],"promise":[3],"to":[4,97],"improve":[5,135],"agents":[6],"by":[7],"turning":[8],"execution":[9],"feedback":[10,66,95,143,172],"into":[11],"persistent":[12,216],"updates":[14],"without":[15],"changing":[16],"the":[17,42,78,94,98,150,187,199],"underlying":[18],"model.":[19],"Yet":[20],"it":[21],"remains":[22,204],"unclear":[23],"when":[24],"further":[25],"evolution":[26],"helps,":[27],"how":[28],"successful":[29],"and":[30,35,58,80,87,101,155,161,167,228],"failed":[31,147],"trajectories":[32],"shape":[33],"revision,":[34],"whether":[36],"extra":[37],"test-time":[38],"computation":[39],"can":[40],"recover":[41],"same":[43],"gains.":[44],"To":[45],"address":[46],"these":[47],"questions,":[48],"we":[49,76],"present":[50],"a":[51],"controlled":[52],"evaluation":[53],"framework":[54],"across":[55,68,158],"five":[56],"benchmarks":[57],"three":[59],"models.":[60],"Our":[61],"primary":[62],"study":[63],"contains":[64],"42":[65],"runs":[67],"14":[69,130],"supported":[70],"model-benchmark":[71],"settings.":[72,159],"Within":[73],"each":[74],"setting,":[75],"hold":[77],"executor":[79],"optimizer":[81],"configuration,":[82],"revision":[83],"procedure,":[84],"validation":[85,119],"rule,":[86],"round":[88],"budget":[89],"fixed,":[90],"while":[91],"varying":[92],"only":[93,112],"shown":[96],"optimizer:":[99],"successes":[100,107],"failures":[102,104],"(Normal),":[103],"only,":[105],"or":[106],"only.":[108],"Evolution":[109],"is":[110,219,240],"sparse:":[111],"55":[113],"of":[114,129,133,153,198],"388":[115],"candidates":[116],"establish":[117],"byte-distinct":[118],"bests.":[120],"Validation-based":[121],"selection":[122],"chooses":[123],"an":[124],"evolved":[125,200],"in":[127],"11":[128,139],"settings,":[131],"nine":[132],"which":[134],"released-test":[136],"performance.":[137],"All":[138],"selections":[140],"come":[141],"from":[142,235],"conditions":[144],"that":[145],"include":[146],"trajectories,":[148],"although":[149],"relative":[151],"ranking":[152],"Normal":[154],"Fail-only":[156],"varies":[157],"Validation":[160],"downstream":[162],"evaluations":[163],"on":[164,208],"test,":[165],"robustness,":[166],"transfer":[168],"sometimes":[169],"favor":[170],"different":[171],"views.":[173],"A":[174],"broader":[175],"SearchQA":[176,201],"analysis":[177],"covering":[178],"eight":[179],"models":[180],"shows":[181],"similarly":[182],"sparse,":[183,223],"feedback-dependent":[184],"dynamics.":[185],"In":[186],"GPT-5.5":[188],"test-time-scaling":[189],"controls,":[190],"oracle":[191],"Parallel":[192],"Sampling":[193],"comes":[194],"within":[195],"0.43":[196],"points":[197,206],"but":[203],"30.96":[205],"behind":[207],"SpreadsheetBench;":[209],"Sequential":[210],"Refinement":[211],"recovers":[212],"neither":[213],"gain.":[214],"Overall,":[215],"self-evolution":[218],"better":[220],"understood":[221],"as":[222],"validation-filtered":[224],"search":[225],"with":[226],"model-":[227],"benchmark-dependent":[229],"returns,":[230],"rather":[231],"than":[232],"steady":[233],"improvement":[234],"additional":[236],"rounds.":[237],"The":[238],"implementation":[239],"available":[241],"at":[242],"https://github.com/HKUST-KnowComp/rethinkskill.":[243]},"counts_by_year":[],"updated_date":"2026-10-28T10:33:24.094632","created_date":"2026-08-06T00:00:00"}
