{"id":"https://openalex.org/W7150914366","doi":"https://doi.org/10.1146/annurev-statistics-042925-123430","title":"Reinforcement Learning from Human Feedback: A Statistical Perspective","display_name":"Reinforcement Learning from Human Feedback: A Statistical Perspective","publication_year":2026,"publication_date":"2026-11-03","ids":{"openalex":"https://openalex.org/W7150914366","doi":"https://doi.org/10.1146/annurev-statistics-042925-123430"},"language":"en","primary_location":{"id":"doi:10.1146/annurev-statistics-042925-123430","doi":"https://doi.org/10.1146/annurev-statistics-042925-123430","is_oa":false,"landing_page_url":"https://doi.org/10.1146/annurev-statistics-042925-123430","pdf_url":null,"source":{"id":"https://openalex.org/S4210217375","display_name":"Annual Review of Statistics and Its Application","issn_l":"2326-8298","issn":["2326-8298","2326-831X"],"is_oa":false,"is_in_doaj":false,"is_core":true,"listed_in":["anvur-scientific","cwts-core","dongbi-b","doyens","jufo-1","ki-jl-1","norway-1","poland-200","russia-white-list-1","vabb-shw"],"host_organization":"https://openalex.org/P4310320373","host_organization_name":"Annual Reviews","host_organization_lineage":["https://openalex.org/P4310320373"],"host_organization_lineage_names":["Annual Reviews"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Annual Review of Statistics and Its Application","raw_type":"journal-article"},"volume":null,"type":"article","indexed_in":["arxiv","crossref","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2604.02507","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5069422514","display_name":"Pangpang Liu","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I32971472","display_name":"Yale University","ror":"https://ror.org/03v76x132","country_code":"US","type":"education","lineage":["https://openalex.org/I32971472"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Pangpang Liu","raw_affiliation_strings":["1Department of Biostatistics, Yale University, New Haven, Connecticut, United States; email:\u00a0pangpang.liu@yale.edu"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"1Department of Biostatistics, Yale University, New Haven, Connecticut, United States; email:\u00a0pangpang.liu@yale.edu","institution_ids":["https://openalex.org/I32971472"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5124456950","display_name":"Chengchun Shi","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I909854389","display_name":"London School of Economics and Political Science","ror":"https://ror.org/0090zs177","country_code":"GB","type":"education","lineage":["https://openalex.org/I124357947","https://openalex.org/I909854389"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Chengchun Shi","raw_affiliation_strings":["2Department of Statistics, London School of Economics and Political Science, London, United Kingdom; email:\u00a0c.shi7@lse.ac.uk"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"2Department of Statistics, London School of Economics and Political Science, London, United Kingdom; email:\u00a0c.shi7@lse.ac.uk","institution_ids":["https://openalex.org/I909854389"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5133045792","display_name":"Will Wei Sun","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I219193219","display_name":"Purdue University West Lafayette","ror":"https://ror.org/02dqehb95","country_code":"US","type":"education","lineage":["https://openalex.org/I219193219"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Will Wei Sun","raw_affiliation_strings":["3Mitch Daniels School of Business, Purdue University, West Lafayette, Indiana, United States; email:\u00a0sun244@purdue.edu"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"3Mitch Daniels School of Business, Purdue University, West Lafayette, Indiana, United States; email:\u00a0sun244@purdue.edu","institution_ids":["https://openalex.org/I219193219"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":3,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.17884108,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"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":0.9291999936103821,"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":0.9291999936103821,"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":0.037700001150369644,"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/T12101","display_name":"Advanced Bandit Algorithms Research","score":0.00559999980032444,"subfield":{"id":"https://openalex.org/subfields/1803","display_name":"Management Science and Operations Research"},"field":{"id":"https://openalex.org/fields/18","display_name":"Decision Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/reinforcement-learning-from-human-feedback","display_name":"reinforcement learning from human feedback","score":0.9950000047683716},{"id":"https://openalex.org/keywords/large-language-model-alignment","display_name":"large language model alignment","score":0.9459999799728394},{"id":"https://openalex.org/keywords/reward-modeling","display_name":"reward modeling","score":0.9559999704360962},{"id":"https://openalex.org/keywords/pairwise-preference-learning","display_name":"pairwise preference learning","score":0.7070000171661377},{"id":"https://openalex.org/keywords/bradley-terry-luce-model","display_name":"Bradley-Terry-Luce model","score":0.9710000157356262},{"id":"https://openalex.org/keywords/active-learning-machine-learning","display_name":"active learning (machine learning)","score":0.9480000138282776},{"id":"https://openalex.org/keywords/experimental-design","display_name":"experimental design","score":0.9890000224113464},{"id":"https://openalex.org/keywords/uncertainty-quantification","display_name":"uncertainty quantification","score":0.9959999918937683},{"id":"https://openalex.org/keywords/direct-preference-optimization","display_name":"direct preference optimization","score":0.8500000238418579},{"id":"https://openalex.org/keywords/reinforcement-learning-from-ai-feedback","display_name":"reinforcement learning from AI feedback","score":0.8709999918937683},{"id":"https://openalex.org/keywords/reinforcement-learning-from-verifiable-rewards","display_name":"reinforcement learning from verifiable rewards","score":0.8930000066757202}],"concepts":[{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.7842000126838684},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.600600004196167},{"id":"https://openalex.org/C12713177","wikidata":"https://www.wikidata.org/wiki/Q1900281","display_name":"Perspective (graphical)","level":2,"score":0.5989999771118164},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5598000288009644},{"id":"https://openalex.org/C184898388","wikidata":"https://www.wikidata.org/wiki/Q1435712","display_name":"Pairwise comparison","level":2,"score":0.512499988079071},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4970000088214874},{"id":"https://openalex.org/C2781249084","wikidata":"https://www.wikidata.org/wiki/Q908656","display_name":"Preference","level":2,"score":0.46860000491142273},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.4097999930381775},{"id":"https://openalex.org/C114289077","wikidata":"https://www.wikidata.org/wiki/Q3284399","display_name":"Statistical model","level":2,"score":0.3824999928474426},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.3662000000476837},{"id":"https://openalex.org/C181204326","wikidata":"https://www.wikidata.org/wiki/Q7239820","display_name":"Preference learning","level":3,"score":0.34459999203681946},{"id":"https://openalex.org/C77967617","wikidata":"https://www.wikidata.org/wiki/Q4677561","display_name":"Active learning (machine learning)","level":2,"score":0.32839998602867126},{"id":"https://openalex.org/C67203356","wikidata":"https://www.wikidata.org/wiki/Q1321905","display_name":"Reinforcement","level":2,"score":0.3244999945163727},{"id":"https://openalex.org/C85847156","wikidata":"https://www.wikidata.org/wiki/Q59015987","display_name":"Verifiable secret sharing","level":3,"score":0.2964000105857849},{"id":"https://openalex.org/C32254414","wikidata":"https://www.wikidata.org/wiki/Q4724364","display_name":"Algorithmic learning theory","level":3,"score":0.25760000944137573}],"mesh":[],"locations_count":3,"locations":[{"id":"doi:10.1146/annurev-statistics-042925-123430","doi":"https://doi.org/10.1146/annurev-statistics-042925-123430","is_oa":false,"landing_page_url":"https://doi.org/10.1146/annurev-statistics-042925-123430","pdf_url":null,"source":{"id":"https://openalex.org/S4210217375","display_name":"Annual Review of Statistics and Its Application","issn_l":"2326-8298","issn":["2326-8298","2326-831X"],"is_oa":false,"is_in_doaj":false,"is_core":true,"listed_in":["anvur-scientific","cwts-core","dongbi-b","doyens","jufo-1","ki-jl-1","norway-1","poland-200","russia-white-list-1","vabb-shw"],"host_organization":"https://openalex.org/P4310320373","host_organization_name":"Annual Reviews","host_organization_lineage":["https://openalex.org/P4310320373"],"host_organization_lineage_names":["Annual Reviews"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Annual Review of Statistics and Its Application","raw_type":"journal-article"},{"id":"pmh:oai:arXiv.org:2604.02507","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2604.02507","pdf_url":"https://arxiv.org/pdf/2604.02507","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.2604.02507","doi":"https://doi.org/10.48550/arxiv.2604.02507","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.02507","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:2604.02507","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2604.02507","pdf_url":"https://arxiv.org/pdf/2604.02507","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":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7150914366.pdf","grobid_xml":"https://content.openalex.org/works/W7150914366.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4408595355","https://openalex.org/W4417026644","https://openalex.org/W4390215524","https://openalex.org/W4387892197","https://openalex.org/W7161787650","https://openalex.org/W4414681422","https://openalex.org/W4417521172","https://openalex.org/W4411063470","https://openalex.org/W4407247462","https://openalex.org/W4414520407"],"abstract_inverted_index":{"Reinforcement":[0],"learning":[1,100,132,141],"from":[2,103,133,142],"human":[3,19],"feedback":[4,38],"(RLHF)":[5],"has":[6],"emerged":[7],"as":[8,82,121,145,147],"a":[9,49],"central":[10],"framework":[11],"for":[12,99,108],"aligning":[13],"large":[14],"language":[15],"models":[16,42],"(LLMs)":[17],"with":[18],"preferences.":[20],"Despite":[21],"its":[22],"practical":[23],"success,":[24],"RLHF":[25,114,157],"raises":[26],"fundamental":[27],"statistical":[28,50,79],"questions":[29],"because":[30],"it":[31],"relies":[32],"on":[33,52,56],"noisy,":[34],"subjective,":[35],"and":[36,43,71,74,93,107,116,139,152],"heterogeneous":[37],"to":[39,77],"learn":[40],"reward":[41,69,101],"optimize":[44],"policies.":[45],"This":[46],"review":[47,97],"provides":[48],"perspective":[51],"RLHF,":[53,65],"focusing":[54],"primarily":[55],"LLM":[57],"alignment.":[58],"We":[59,96,125,159],"introduce":[60],"the":[61,83],"main":[62],"components":[63],"of":[64],"including":[66,130],"supervised":[67],"fine-tuning,":[68],"modeling,":[70],"policy":[72],"optimization,":[73],"relate":[75],"them":[76],"familiar":[78],"concepts":[80],"such":[81,120],"Bradley\u2013Terry\u2013Luce":[84],"model,":[85],"latent":[86],"utility":[87],"estimation,":[88],"active":[89],"learning,":[90],"experimental":[91],"design,":[92],"uncertainty":[94],"quantification.":[95],"methods":[98],"functions":[102],"pairwise":[104],"preference":[105,123],"data":[106],"optimizing":[109],"policies":[110],"through":[111],"both":[112],"two-stage":[113],"pipelines":[115],"emerging":[117],"one-stage":[118],"approaches":[119],"direct":[122],"optimization.":[124],"further":[126],"discuss":[127],"recent":[128],"extensions":[129],"reinforcement":[131,140],"artificial":[134],"intelligence":[135],"feedback,":[136],"inference-time":[137],"algorithms,":[138],"verifiable":[143],"rewards,":[144],"well":[146],"benchmark":[148],"datasets,":[149],"evaluation":[150],"protocols,":[151],"open-source":[153],"frameworks":[154],"that":[155],"support":[156],"research.":[158],"conclude":[160],"by":[161],"highlighting":[162],"open":[163],"challenges":[164],"in":[165],"RLHF.":[166]},"counts_by_year":[],"updated_date":"2026-11-05T12:25:35.769452","created_date":"2026-04-07T00:00:00"}