{"id":"https://openalex.org/W7224908043","doi":"https://doi.org/10.48550/arxiv.2610.38637","title":"On-Policy Self-Distillation with Teacher Evolution","display_name":"On-Policy Self-Distillation with Teacher Evolution","publication_year":2026,"publication_date":"2026-10-28","ids":{"openalex":"https://openalex.org/W7224908043","doi":"https://doi.org/10.48550/arxiv.2610.38637"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2610.38637","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.38637","pdf_url":"https://arxiv.org/pdf/2610.38637","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-nc-sa","license_id":"https://openalex.org/licenses/cc-by-nc-sa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2610.38637","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5040880537","display_name":"Shufan Shen","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I19820366","display_name":"Chinese Academy of Sciences","ror":"https://ror.org/034t30j35","country_code":"CN","type":"government","lineage":["https://openalex.org/I19820366"]},{"id":"https://openalex.org/I4210090176","display_name":"Institute of Computing Technology","ror":"https://ror.org/0090r4d87","country_code":"CN","type":"facility","lineage":["https://openalex.org/I19820366","https://openalex.org/I4210090176"]},{"id":"https://openalex.org/I4210165038","display_name":"University of Chinese Academy of Sciences","ror":"https://ror.org/05qbk4x57","country_code":"CN","type":"education","lineage":["https://openalex.org/I19820366","https://openalex.org/I4210165038"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Shen, Shufan","raw_affiliation_strings":["State Key Lab. of AI Safety , Institute of Computing Technology , Chinese Academy of Sciences","University of Chinese Academy of Sciences"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"State Key Lab. of AI Safety , Institute of Computing Technology , Chinese Academy of Sciences","institution_ids":["https://openalex.org/I19820366","https://openalex.org/I4210090176"]},{"raw_affiliation_string":"University of Chinese Academy of Sciences","institution_ids":["https://openalex.org/I4210165038"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5017128773","display_name":"Zhongni Hou","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I4405259410","display_name":"Meituan (China)","ror":"https://ror.org/0272kpg92","country_code":"CN","type":"company","lineage":["https://openalex.org/I4405259410"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Hou, Zhongni","raw_affiliation_strings":["Meituan"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meituan","institution_ids":["https://openalex.org/I4405259410"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129918591","display_name":"Chao Bi","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I19820366","display_name":"Chinese Academy of Sciences","ror":"https://ror.org/034t30j35","country_code":"CN","type":"government","lineage":["https://openalex.org/I19820366"]},{"id":"https://openalex.org/I4210090176","display_name":"Institute of Computing Technology","ror":"https://ror.org/0090r4d87","country_code":"CN","type":"facility","lineage":["https://openalex.org/I19820366","https://openalex.org/I4210090176"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Bi, Chao","raw_affiliation_strings":["State Key Lab. of AI Safety , Institute of Computing Technology , Chinese Academy of Sciences"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"State Key Lab. of AI Safety , Institute of Computing Technology , Chinese Academy of Sciences","institution_ids":["https://openalex.org/I19820366","https://openalex.org/I4210090176"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5160230472","display_name":"Yufei Zhang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I4405259410","display_name":"Meituan (China)","ror":"https://ror.org/0272kpg92","country_code":"CN","type":"company","lineage":["https://openalex.org/I4405259410"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhang, Yufei","raw_affiliation_strings":["Meituan"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meituan","institution_ids":["https://openalex.org/I4405259410"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5033440540","display_name":"Guojun Yin","orcid":"https://orcid.org/0000-0001-8260-5449","observed_orcids":["https://orcid.org/0000-0001-8260-5449"]},"institutions":[{"id":"https://openalex.org/I4405259410","display_name":"Meituan (China)","ror":"https://ror.org/0272kpg92","country_code":"CN","type":"company","lineage":["https://openalex.org/I4405259410"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yin, Guojun","raw_affiliation_strings":["Meituan"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meituan","institution_ids":["https://openalex.org/I4405259410"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5160232099","display_name":"Shuhui Wang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I19820366","display_name":"Chinese Academy of Sciences","ror":"https://ror.org/034t30j35","country_code":"CN","type":"government","lineage":["https://openalex.org/I19820366"]},{"id":"https://openalex.org/I4210090176","display_name":"Institute of Computing Technology","ror":"https://ror.org/0090r4d87","country_code":"CN","type":"facility","lineage":["https://openalex.org/I19820366","https://openalex.org/I4210090176"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Wang, Shuhui","raw_affiliation_strings":["State Key Lab. of AI Safety , Institute of Computing Technology , Chinese Academy of Sciences"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"State Key Lab. of AI Safety , Institute of Computing Technology , Chinese Academy of Sciences","institution_ids":["https://openalex.org/I19820366","https://openalex.org/I4210090176"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":4,"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/T10028","display_name":"Topic Modeling","score":0.5526999831199646,"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":0.5526999831199646,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.16349999606609344,"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":0.08780000358819962,"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/on-policy-self-distillation","display_name":"on-policy self-distillation","score":0.9760000109672546},{"id":"https://openalex.org/keywords/reward-sparsity","display_name":"reward sparsity","score":0.8159999847412109},{"id":"https://openalex.org/keywords/post-training","display_name":"post-training","score":0.5},{"id":"https://openalex.org/keywords/reinforcement-learning-from-human-feedback","display_name":"reinforcement learning from human feedback","score":0.6290000081062317}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6937000155448914},{"id":"https://openalex.org/C55439883","wikidata":"https://www.wikidata.org/wiki/Q360812","display_name":"Correctness","level":2,"score":0.682699978351593},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.6348000168800354},{"id":"https://openalex.org/C43214815","wikidata":"https://www.wikidata.org/wiki/Q7310987","display_name":"Reliability (semiconductor)","level":3,"score":0.6262999773025513},{"id":"https://openalex.org/C2779530757","wikidata":"https://www.wikidata.org/wiki/Q1207505","display_name":"Quality (philosophy)","level":2,"score":0.5957000255584717},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.5928999781608582},{"id":"https://openalex.org/C112930515","wikidata":"https://www.wikidata.org/wiki/Q4389547","display_name":"Risk analysis (engineering)","level":1,"score":0.35100001096725464},{"id":"https://openalex.org/C50335755","wikidata":"https://www.wikidata.org/wiki/Q483247","display_name":"Phenomenon","level":2,"score":0.33820000290870667},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.32280001044273376},{"id":"https://openalex.org/C2778755073","wikidata":"https://www.wikidata.org/wiki/Q10858537","display_name":"Scale (ratio)","level":2,"score":0.3206999897956848},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3131999969482422},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.29809999465942383},{"id":"https://openalex.org/C110121322","wikidata":"https://www.wikidata.org/wiki/Q865811","display_name":"Distribution (mathematics)","level":2,"score":0.26980000734329224},{"id":"https://openalex.org/C539667460","wikidata":"https://www.wikidata.org/wiki/Q2414942","display_name":"Management science","level":1,"score":0.267300009727478},{"id":"https://openalex.org/C2780009758","wikidata":"https://www.wikidata.org/wiki/Q6804172","display_name":"Measure (data warehouse)","level":2,"score":0.26409998536109924},{"id":"https://openalex.org/C2775924081","wikidata":"https://www.wikidata.org/wiki/Q55608371","display_name":"Control (management)","level":2,"score":0.2623000144958496},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.2621999979019165},{"id":"https://openalex.org/C195094911","wikidata":"https://www.wikidata.org/wiki/Q14167904","display_name":"Process management","level":1,"score":0.2526000142097473}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2610.38637","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.38637","pdf_url":"https://arxiv.org/pdf/2610.38637","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-nc-sa","license_id":"https://openalex.org/licenses/cc-by-nc-sa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2610.38637","doi":"https://doi.org/10.48550/arxiv.2610.38637","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2610.38637","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-nc-sa","license_id":"https://openalex.org/licenses/cc-by-nc-sa","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2610.38637","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.38637","pdf_url":"https://arxiv.org/pdf/2610.38637","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-nc-sa","license_id":"https://openalex.org/licenses/cc-by-nc-sa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[{"display_name":"Quality education","id":"https://metadata.un.org/sdg/4","score":0.8277000188827515}],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7224908043.pdf","grobid_xml":"https://content.openalex.org/works/W7224908043.grobid-xml"},"referenced_works_count":28,"referenced_works":["https://openalex.org/W2794325560","https://openalex.org/W3177813494","https://openalex.org/W3210904070","https://openalex.org/W4327810158","https://openalex.org/W4380136738","https://openalex.org/W4385970164","https://openalex.org/W4388718089","https://openalex.org/W4391631327","https://openalex.org/W4399695623","https://openalex.org/W4403883066","https://openalex.org/W4409049124","https://openalex.org/W4410312401","https://openalex.org/W4414281281","https://openalex.org/W4414939355","https://openalex.org/W4415277433","https://openalex.org/W4417305769","https://openalex.org/W4417465430","https://openalex.org/W6929370958","https://openalex.org/W7125861034","https://openalex.org/W7125915072","https://openalex.org/W7126034091","https://openalex.org/W7130402983","https://openalex.org/W7134813685","https://openalex.org/W7150828934","https://openalex.org/W7154505894","https://openalex.org/W7161162008","https://openalex.org/W7164364162","https://openalex.org/W7214984162"],"related_works":["https://openalex.org/W7204449334","https://openalex.org/W7202207540","https://openalex.org/W7204908525","https://openalex.org/W7214670177","https://openalex.org/W7220652251","https://openalex.org/W7202177007","https://openalex.org/W7214887446","https://openalex.org/W7222997000","https://openalex.org/W7224847320","https://openalex.org/W7160513348"],"abstract_inverted_index":{"On-Policy":[0],"Self-Distillation":[1,72],"(OPSD),":[2],"which":[3,74],"aligns":[4],"a":[5,22],"student":[6],"model's":[7,51,107],"token-level":[8,146],"distribution":[9],"with":[10,129],"that":[11,49,141],"of":[12,153],"its":[13],"self-teacher":[14,77],"prompted":[15],"by":[16,58,113],"privileged":[17],"context,":[18],"has":[19],"emerged":[20],"as":[21,96],"promising":[23],"direction":[24],"to":[25,37,78,103,126],"mitigate":[26],"reward":[27],"sparsity":[28],"in":[29,157],"sequence-level":[30,148],"post-training":[31],"paradigms.":[32],"However,":[33],"OPSD":[34],"often":[35],"fails":[36],"achieve":[38],"continuous":[39],"performance":[40],"gains":[41],"during":[42,110],"training.":[43],"In":[44],"this":[45,56,67],"paper,":[46],"we":[47,69,99,121],"find":[48],"the":[50,76,106,151],"limited":[52],"self-teaching":[53,108],"capability":[54,109],"underlies":[55],"phenomenon":[57],"generating":[59],"unreliable":[60,130],"guidance":[61,116],"on":[62,86],"certain":[63],"samples.":[64],"To":[65],"address":[66],"problem,":[68],"propose":[70],"Teacher-Evolution":[71],"(TESD),":[73],"enables":[75],"generate":[79],"rollouts":[80],"and":[81,89,137,147],"receive":[82],"environmental":[83],"feedback":[84],"based":[85],"task":[87],"correctness":[88],"context":[90],"independence.":[91],"First,":[92],"using":[93],"these":[94,119],"rewards":[95],"training":[97],"objectives,":[98],"perform":[100],"teacher-oriented":[101],"optimization":[102,125],"continuously":[104],"enhance":[105],"distillation.":[111],"Second,":[112],"estimating":[114],"teacher":[115],"reliability":[117],"from":[118],"rewards,":[120],"apply":[122],"adaptive":[123],"student-oriented":[124],"down-weight":[127],"samples":[128],"guidance.":[131],"Extensive":[132],"experiments":[133],"across":[134],"multiple":[135],"datasets":[136],"model":[138],"scales":[139],"show":[140],"TESD":[142],"consistently":[143],"outperforms":[144],"representative":[145],"methods,":[149],"highlighting":[150],"necessity":[152],"improving":[154],"supervision":[155],"quality":[156],"OPSD.":[158]},"counts_by_year":[],"updated_date":"2026-11-10T11:39:43.021295","created_date":"2026-10-31T00:00:00"}