{"id":"https://openalex.org/W7161939691","doi":"https://doi.org/10.48550/arxiv.2605.20740","title":"Distribution-Aware Reward: Reinforcement Learning over Predictive Distributions for LLM Regression","display_name":"Distribution-Aware Reward: Reinforcement Learning over Predictive Distributions for LLM Regression","publication_year":2026,"publication_date":"2026-05-20","ids":{"openalex":"https://openalex.org/W7161939691","doi":"https://doi.org/10.48550/arxiv.2605.20740"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2605.20740","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.20740","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,"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"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2605.20740","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5136645480","display_name":"Jungsoo Park","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Park, Jungsoo","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5087248105","display_name":"Hyungjoo Chae","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chae, Hyungjoo","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136623221","display_name":"Ethan Mendes","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Mendes, Ethan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136641348","display_name":"Jay DeYoung","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"DeYoung, Jay","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136645449","display_name":"Varsha Kishore","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Kishore, Varsha","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136668806","display_name":"Dr Wei Xu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xu, Wei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5136725257","display_name":"Alan Ritter","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ritter, Alan","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":false,"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/T11948","display_name":"Machine Learning in Materials Science","score":0.44339999556541443,"subfield":{"id":"https://openalex.org/subfields/2505","display_name":"Materials Chemistry"},"field":{"id":"https://openalex.org/fields/25","display_name":"Materials Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T11948","display_name":"Machine Learning in Materials Science","score":0.44339999556541443,"subfield":{"id":"https://openalex.org/subfields/2505","display_name":"Materials Chemistry"},"field":{"id":"https://openalex.org/fields/25","display_name":"Materials Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.10570000112056732,"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/T12535","display_name":"Machine Learning and Data Classification","score":0.061400000005960464,"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/regression","display_name":"Regression","score":0.6061999797821045},{"id":"https://openalex.org/keywords/reinforcement-learning","display_name":"Reinforcement learning","score":0.5217000246047974},{"id":"https://openalex.org/keywords/ranking","display_name":"Ranking (information retrieval)","score":0.5130000114440918},{"id":"https://openalex.org/keywords/pointwise","display_name":"Pointwise","score":0.49070000648498535},{"id":"https://openalex.org/keywords/property","display_name":"Property (philosophy)","score":0.396699994802475},{"id":"https://openalex.org/keywords/code","display_name":"Code (set theory)","score":0.39660000801086426},{"id":"https://openalex.org/keywords/probability-distribution","display_name":"Probability distribution","score":0.3919999897480011},{"id":"https://openalex.org/keywords/regression-analysis","display_name":"Regression analysis","score":0.38749998807907104}],"concepts":[{"id":"https://openalex.org/C83546350","wikidata":"https://www.wikidata.org/wiki/Q1139051","display_name":"Regression","level":2,"score":0.6061999797821045},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5734999775886536},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5267000198364258},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.5217000246047974},{"id":"https://openalex.org/C189430467","wikidata":"https://www.wikidata.org/wiki/Q7293293","display_name":"Ranking (information retrieval)","level":2,"score":0.5130000114440918},{"id":"https://openalex.org/C2777984123","wikidata":"https://www.wikidata.org/wiki/Q9248237","display_name":"Pointwise","level":2,"score":0.49070000648498535},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.487199991941452},{"id":"https://openalex.org/C189950617","wikidata":"https://www.wikidata.org/wiki/Q937228","display_name":"Property (philosophy)","level":2,"score":0.396699994802475},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.39660000801086426},{"id":"https://openalex.org/C149441793","wikidata":"https://www.wikidata.org/wiki/Q200726","display_name":"Probability distribution","level":2,"score":0.3919999897480011},{"id":"https://openalex.org/C152877465","wikidata":"https://www.wikidata.org/wiki/Q208042","display_name":"Regression analysis","level":2,"score":0.38749998807907104},{"id":"https://openalex.org/C45804977","wikidata":"https://www.wikidata.org/wiki/Q7239673","display_name":"Predictive modelling","level":2,"score":0.3269999921321869},{"id":"https://openalex.org/C172205157","wikidata":"https://www.wikidata.org/wiki/Q1782962","display_name":"Model predictive control","level":3,"score":0.32280001044273376},{"id":"https://openalex.org/C117220453","wikidata":"https://www.wikidata.org/wiki/Q5172842","display_name":"Correlation","level":2,"score":0.32019999623298645},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.3172000050544739},{"id":"https://openalex.org/C43126263","wikidata":"https://www.wikidata.org/wiki/Q128751","display_name":"Source code","level":2,"score":0.3059999942779541},{"id":"https://openalex.org/C105795698","wikidata":"https://www.wikidata.org/wiki/Q12483","display_name":"Statistics","level":1,"score":0.30410000681877136},{"id":"https://openalex.org/C28719098","wikidata":"https://www.wikidata.org/wiki/Q44946","display_name":"Point (geometry)","level":2,"score":0.3012999892234802},{"id":"https://openalex.org/C48921125","wikidata":"https://www.wikidata.org/wiki/Q10861030","display_name":"Linear regression","level":2,"score":0.2906999886035919},{"id":"https://openalex.org/C177769412","wikidata":"https://www.wikidata.org/wiki/Q278090","display_name":"Prior probability","level":3,"score":0.2766999900341034},{"id":"https://openalex.org/C32230216","wikidata":"https://www.wikidata.org/wiki/Q7882499","display_name":"Uncertainty quantification","level":2,"score":0.2754000127315521},{"id":"https://openalex.org/C67203356","wikidata":"https://www.wikidata.org/wiki/Q1321905","display_name":"Reinforcement","level":2,"score":0.2635999917984009},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.2606000006198883},{"id":"https://openalex.org/C136389625","wikidata":"https://www.wikidata.org/wiki/Q334384","display_name":"Supervised learning","level":3,"score":0.2540999948978424}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2605.20740","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.20740","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,"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":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2605.20740","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.20740","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,"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"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Large":[0],"language":[1,59,201],"models":[2,60],"can":[3],"predict":[4],"real-valued":[5],"quantities":[6],"from":[7,136],"heterogeneous":[8],"inputs":[9],"such":[10],"as":[11,85],"text,":[12],"code,":[13],"and":[14,98,117,132,147,176,190,206],"molecular":[15,133,178],"strings,":[16],"but":[17],"most":[18],"training":[19],"objectives":[20],"score":[21],"each":[22,104],"decoded":[23,74,83],"floating-point":[24],"number":[25],"independently,":[26],"improving":[27],"point":[28],"estimates":[29],"without":[30],"ensuring":[31],"calibrated":[32],"predictive":[33,64,88,198],"distributions.":[34],"This":[35],"limits":[36],"applications":[37],"requiring":[38],"candidate":[39],"ranking":[40],"or":[41],"uncertainty":[42,192],"estimation.":[43],"We":[44,120],"introduce":[45],"Distribution-Aware":[46],"Reward,":[47],"an":[48,86],"on-policy":[49],"reinforcement":[50,149],"learning":[51,150],"objective":[52],"whose":[53],"main":[54],"contribution":[55,107],"is":[56],"to":[57,61,108],"train":[58],"produce":[62],"better":[63,207],"distributions":[65,199],"for":[66],"regression":[67,203],"tasks,":[68,140],"rather":[69],"than":[70],"only":[71,167],"optimizing":[72,197],"individual":[73],"outputs":[75],"against":[76],"scalar":[77],"targets.":[78],"Our":[79],"method":[80,123,142,185],"treats":[81],"multiple":[82],"samples":[84],"empirical":[87],"distribution,":[89],"evaluates":[90],"it":[91,165],"with":[92,152,173],"the":[93],"Continuous":[94],"Ranked":[95],"Probability":[96],"Score,":[97],"assigns":[99],"leave-one-out":[100],"credit":[101],"based":[102],"on":[103,124,161],"rollout's":[105],"marginal":[106],"distribution":[109],"quality,":[110],"rewarding":[111],"predictions":[112],"that":[113,183,195],"are":[114],"both":[115],"accurate":[116],"appropriately":[118],"dispersed.":[119],"evaluate":[121],"our":[122,141,184],"a":[125,157],"controlled":[126],"Gaussian-mixture":[127],"task,":[128],"code":[129],"performance":[130],"prediction,":[131],"property":[134],"prediction":[135],"SMILES":[137,168],"strings.":[138],"Across":[139],"improves":[143,191],"over":[144],"supervised":[145],"fine-tuning":[146],"pointwise":[148],"baselines,":[151],"strong":[153,174],"rank-correlation":[154],"gains,":[155],"including":[156],"6-point":[158],"Spearman":[159],"improvement":[160],"KBSS.":[162],"On":[163],"MoleculeNet,":[164],"uses":[166],"strings":[169],"yet":[170],"remains":[171],"competitive":[172],"graph-based":[175],"3D":[177],"models.":[179],"Further":[180],"analyses":[181],"show":[182],"mitigates":[186],"rollout":[187],"diversity":[188],"collapse":[189],"diagnostics,":[193],"suggesting":[194],"directly":[196],"makes":[200],"model":[202],"more":[204],"robust":[205],"calibrated.":[208]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2026-05-22T00:00:00"}