{"id":"https://openalex.org/W7214901732","doi":"https://doi.org/10.48550/arxiv.2609.36535","title":"When Updating Stops Being Learning: Rethinking LLM Self-Evolution via learnable information gain","display_name":"When Updating Stops Being Learning: Rethinking LLM Self-Evolution via learnable information gain","publication_year":2026,"publication_date":"2026-09-29","ids":{"openalex":"https://openalex.org/W7214901732","doi":"https://doi.org/10.48550/arxiv.2609.36535"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2609.36535","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.36535","pdf_url":"https://arxiv.org/pdf/2609.36535","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","is_traditional":true,"indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2609.36535","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5153613881","display_name":"Chenxu Wang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I139759216","display_name":"Beijing University of Posts and Telecommunications","ror":"https://ror.org/04w9fbh59","country_code":"CN","type":"education","lineage":["https://openalex.org/I139759216"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Wang, Chenxu","raw_affiliation_strings":["Beijing University of Posts and Telecommunications"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Beijing University of Posts and Telecommunications","institution_ids":["https://openalex.org/I139759216"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5153584821","display_name":"Chaozhuo Li","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I4210100255","display_name":"Beijing Academy of Artificial Intelligence","ror":"https://ror.org/016a74861","country_code":"CN","type":"other","lineage":["https://openalex.org/I4210100255"]},{"id":"https://openalex.org/I4210116924","display_name":"Chinese University of Hong Kong, Shenzhen","ror":"https://ror.org/02d5ks197","country_code":"CN","type":"education","lineage":["https://openalex.org/I177725633","https://openalex.org/I180726961","https://openalex.org/I4210116924"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Li, Chaozhuo","raw_affiliation_strings":["Beijing Academy of Artificial Intelligence","The Chinese University of Hong Kong , Shenzhen"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Beijing Academy of Artificial Intelligence","institution_ids":["https://openalex.org/I4210100255"]},{"raw_affiliation_string":"The Chinese University of Hong Kong , Shenzhen","institution_ids":["https://openalex.org/I4210116924"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5153213843","display_name":"Xinze Shi","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I139759216","display_name":"Beijing University of Posts and Telecommunications","ror":"https://ror.org/04w9fbh59","country_code":"CN","type":"education","lineage":["https://openalex.org/I139759216"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Shi, Xinze","raw_affiliation_strings":["Beijing University of Posts and Telecommunications"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Beijing University of Posts and Telecommunications","institution_ids":["https://openalex.org/I139759216"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5122097570","display_name":"Songyang Liu","orcid":"https://orcid.org/0009-0003-9592-8046","observed_orcids":["https://orcid.org/0009-0003-9592-8046"]},"institutions":[{"id":"https://openalex.org/I139759216","display_name":"Beijing University of Posts and Telecommunications","ror":"https://ror.org/04w9fbh59","country_code":"CN","type":"education","lineage":["https://openalex.org/I139759216"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Liu, Songyang","raw_affiliation_strings":["Beijing University of Posts and Telecommunications"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Beijing University of Posts and Telecommunications","institution_ids":["https://openalex.org/I139759216"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5153600154","display_name":"Kyrie You Wu","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I139759216","display_name":"Beijing University of Posts and Telecommunications","ror":"https://ror.org/04w9fbh59","country_code":"CN","type":"education","lineage":["https://openalex.org/I139759216"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Wu, Kyrie You","raw_affiliation_strings":["Beijing University of Posts and Telecommunications"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Beijing University of Posts and Telecommunications","institution_ids":["https://openalex.org/I139759216"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5153639598","display_name":"Ziluowen Luo","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I139660479","display_name":"Central South University","ror":"https://ror.org/00f1zfq44","country_code":"CN","type":"education","lineage":["https://openalex.org/I139660479"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Luo, Ziluowen","raw_affiliation_strings":["Central South University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Central South University","institution_ids":["https://openalex.org/I139660479"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5153683333","display_name":"Shun Zhang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I2801345734","display_name":"China Academy of Engineering Physics","ror":"https://ror.org/039vqpp67","country_code":"CN","type":"facility","lineage":["https://openalex.org/I2801345734"]},{"id":"https://openalex.org/I4405256023","display_name":"Graduate School of China Academy of Engineering Physics","ror":"https://ror.org/03p41qy95","country_code":"CN","type":"education","lineage":["https://openalex.org/I4405256023"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhang, Shun","raw_affiliation_strings":["Graduate School of China Academy of Engineering Physics"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Graduate School of China Academy of Engineering Physics","institution_ids":["https://openalex.org/I2801345734","https://openalex.org/I4405256023"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5153597320","display_name":"Chenxi Li","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I4210100255","display_name":"Beijing Academy of Artificial Intelligence","ror":"https://ror.org/016a74861","country_code":"CN","type":"other","lineage":["https://openalex.org/I4210100255"]},{"id":"https://openalex.org/I4210116924","display_name":"Chinese University of Hong Kong, Shenzhen","ror":"https://ror.org/02d5ks197","country_code":"CN","type":"education","lineage":["https://openalex.org/I177725633","https://openalex.org/I180726961","https://openalex.org/I4210116924"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Li, Chenxi","raw_affiliation_strings":["Beijing Academy of Artificial Intelligence","The Chinese University of Hong Kong , Shenzhen"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Beijing Academy of Artificial Intelligence","institution_ids":["https://openalex.org/I4210100255"]},{"raw_affiliation_string":"The Chinese University of Hong Kong , Shenzhen","institution_ids":["https://openalex.org/I4210116924"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5153564332","display_name":"Litian Zhang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I139759216","display_name":"Beijing University of Posts and Telecommunications","ror":"https://ror.org/04w9fbh59","country_code":"CN","type":"education","lineage":["https://openalex.org/I139759216"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhang, Litian","raw_affiliation_strings":["Beijing University of Posts and Telecommunications"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Beijing University of Posts and Telecommunications","institution_ids":["https://openalex.org/I139759216"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":6,"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":0.8574000000953674,"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":0.8574000000953674,"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":0.048700001090765,"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/T12072","display_name":"Machine Learning and Algorithms","score":0.02410000003874302,"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/self-evolution","display_name":"self-evolution","score":0.7289999723434448},{"id":"https://openalex.org/keywords/large-language-models","display_name":"large language models","score":0.9649999737739563},{"id":"https://openalex.org/keywords/self-training","display_name":"self-training","score":0.5180000066757202},{"id":"https://openalex.org/keywords/kullback-leibler-divergence","display_name":"Kullback-Leibler divergence","score":0.9760000109672546},{"id":"https://openalex.org/keywords/sample-reweighting","display_name":"sample reweighting","score":0.8050000071525574},{"id":"https://openalex.org/keywords/atri","display_name":"Atri","score":0.5109999775886536}],"concepts":[{"id":"https://openalex.org/C2983203078","wikidata":"https://www.wikidata.org/wiki/Q255166","display_name":"Information gain","level":2,"score":0.7997999787330627},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6175000071525574},{"id":"https://openalex.org/C106301342","wikidata":"https://www.wikidata.org/wiki/Q4117933","display_name":"Entropy (arrow of time)","level":2,"score":0.5889000296592712},{"id":"https://openalex.org/C207390915","wikidata":"https://www.wikidata.org/wiki/Q1230525","display_name":"Divergence (linguistics)","level":2,"score":0.5453000068664551},{"id":"https://openalex.org/C171752962","wikidata":"https://www.wikidata.org/wiki/Q255166","display_name":"Kullback\u2013Leibler divergence","level":2,"score":0.5411999821662903},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5394999980926514},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.42910000681877136},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.4212000072002411},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3946000039577484},{"id":"https://openalex.org/C101721835","wikidata":"https://www.wikidata.org/wiki/Q813908","display_name":"Conditional entropy","level":3,"score":0.3644999861717224},{"id":"https://openalex.org/C9679016","wikidata":"https://www.wikidata.org/wiki/Q1417473","display_name":"Principle of maximum entropy","level":2,"score":0.3458999991416931},{"id":"https://openalex.org/C168167062","wikidata":"https://www.wikidata.org/wiki/Q1117970","display_name":"Component (thermodynamics)","level":2,"score":0.3248000144958496},{"id":"https://openalex.org/C52622258","wikidata":"https://www.wikidata.org/wiki/Q131222","display_name":"Information theory","level":2,"score":0.30720001459121704},{"id":"https://openalex.org/C2781170535","wikidata":"https://www.wikidata.org/wiki/Q30587856","display_name":"Noisy data","level":2,"score":0.30559998750686646},{"id":"https://openalex.org/C167981619","wikidata":"https://www.wikidata.org/wiki/Q1685498","display_name":"Cross entropy","level":3,"score":0.2822999954223633},{"id":"https://openalex.org/C3020402766","wikidata":"https://www.wikidata.org/wiki/Q104376712","display_name":"Prior information","level":2,"score":0.28209999203681946},{"id":"https://openalex.org/C2780009758","wikidata":"https://www.wikidata.org/wiki/Q6804172","display_name":"Measure (data warehouse)","level":2,"score":0.27959999442100525},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.274399995803833}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2609.36535","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.36535","pdf_url":"https://arxiv.org/pdf/2609.36535","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.2609.36535","doi":"https://doi.org/10.48550/arxiv.2609.36535","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2609.36535","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:2609.36535","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.36535","pdf_url":"https://arxiv.org/pdf/2609.36535","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/W7214901732.pdf","grobid_xml":"https://content.openalex.org/works/W7214901732.grobid-xml"},"referenced_works_count":30,"referenced_works":["https://openalex.org/W3210904070","https://openalex.org/W4386080611","https://openalex.org/W4390573302","https://openalex.org/W4391047180","https://openalex.org/W4391631327","https://openalex.org/W4391800729","https://openalex.org/W4394672701","https://openalex.org/W4400480041","https://openalex.org/W4403964227","https://openalex.org/W4405035967","https://openalex.org/W4407814496","https://openalex.org/W4414683736","https://openalex.org/W4415037569","https://openalex.org/W4415261756","https://openalex.org/W4416010183","https://openalex.org/W4416048648","https://openalex.org/W4417167303","https://openalex.org/W6929629322","https://openalex.org/W7118612177","https://openalex.org/W7128621704","https://openalex.org/W7129067852","https://openalex.org/W7131702487","https://openalex.org/W7155356183","https://openalex.org/W7160951781","https://openalex.org/W7162075892","https://openalex.org/W7162284552","https://openalex.org/W7162416026","https://openalex.org/W7165628670","https://openalex.org/W7172413134","https://openalex.org/W7196963086"],"related_works":["https://openalex.org/W7129088165","https://openalex.org/W7214414372","https://openalex.org/W4407385116","https://openalex.org/W7220713001","https://openalex.org/W4387323597","https://openalex.org/W4416369341","https://openalex.org/W4416845049","https://openalex.org/W7139926871","https://openalex.org/W4395065873","https://openalex.org/W7163212087"],"abstract_inverted_index":{"Self-evolution":[0],"lets":[1],"large":[2],"language":[3,99],"models":[4],"(LLMs)":[5],"improve":[6],"iteratively":[7],"using":[8],"their":[9,88],"own":[10],"generated":[11],"data,":[12],"but":[13],"often":[14],"suffers":[15],"from":[16],"self-evolution":[17,43],"degeneration:":[18],"performance":[19],"improves,":[20],"plateaus,":[21],"then":[22],"declines.":[23],"Existing":[24],"methods":[25],"address":[26],"this":[27,75,114],"issue":[28],"at":[29],"the":[30,35,38,71,78,82,102,145],"component":[31],"level,":[32],"targeting":[33],"either":[34],"Questioner":[36],"or":[37],"Solver,":[39],"and":[40,105,130],"overlook":[41],"that":[42],"is":[44,93],"a":[45,51,66,97,128],"tightly":[46],"coupled":[47],"system.":[48],"We":[49],"propose":[50,117],"holistic":[52],"framework":[53],"based":[54],"on":[55,113,141],"learnable":[56],"information":[57,65,136],"gain,":[58],"which":[59,124],"measures":[60],"how":[61],"much":[62],"novel,":[63],"parameterizable":[64],"round":[67,104,129],"provides":[68],"relative":[69],"to":[70,101],"previous":[72,103],"round.":[73],"Theoretically,":[74],"gain":[76,137],"equals":[77],"Kullback-Leibler":[79],"divergence":[80],"between":[81],"two":[83],"rounds'":[84],"data":[85,108],"distributions":[86],"plus":[87],"entropy":[89],"change.":[90],"Practically,":[91],"it":[92],"estimated":[94],"by":[95],"fitting":[96],"small":[98],"model":[100],"scoring":[106],"new":[107],"via":[109,122],"negative":[110],"log-likelihood.":[111],"Based":[112],"diagnostic,":[115],"we":[116],"ATRI":[118],"(Adaptive":[119],"Training":[120],"Regulation":[121],"Information-gain),":[123],"reweights":[125],"samples":[126],"within":[127],"halts":[131],"training":[132],"across":[133],"rounds":[134],"when":[135],"remains":[138],"low.":[139],"Experiments":[140],"popular":[142],"datasets":[143],"demonstrate":[144],"superiority":[146],"of":[147],"our":[148],"proposal.":[149]},"counts_by_year":[],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-10-01T00:00:00"}