{"id":"https://openalex.org/W7204018933","doi":null,"title":"TreePO: Enhancing Policy Efficacy and Inference Efficiency with Tree Modeling","display_name":"TreePO: Enhancing Policy Efficacy and Inference Efficiency with Tree Modeling","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7204018933"},"language":"en","primary_location":{"id":"pmh:oai:repository.hkust.edu.hk:1783.1-173863","doi":null,"is_oa":false,"landing_page_url":"http://repository.hkust.edu.hk/ir/Record/1783.1-173863","pdf_url":null,"source":{"id":"https://openalex.org/S4306401796","display_name":"Rare & Special e-Zone (The Hong Kong University of Science and Technology)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I200769079","host_organization_name":"Hong Kong University of Science and Technology","host_organization_lineage":["https://openalex.org/I200769079"],"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":"Conference paper"},"volume":null,"type":"conference-paper","indexed_in":[],"open_access":{"is_oa":false,"oa_status":"closed","oa_url":null,"any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5121752602","display_name":"Yizhi Li","orcid":"https://orcid.org/0000-0002-3932-9706","observed_orcids":["https://orcid.org/0000-0002-3932-9706"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Yizhi; id_orcid 0000-0002-3932-9706","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5041193080","display_name":"Qingshui Gu","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Gu, Qingshui","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5147871220","display_name":"Zhoufutu Wen","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wen, Zhoufutu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5147822809","display_name":"Ziniu Li","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Ziniu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5067150934","display_name":"Ruibin Yuan","orcid":"https://orcid.org/0009-0002-0539-6916","observed_orcids":["https://orcid.org/0009-0002-0539-6916"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yuan, Ruibin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5147837007","display_name":"Tianshun Xing","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xing, Tianshun","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5104280301","display_name":"Shuyue Guo","orcid":"https://orcid.org/0000-0002-4780-2794","observed_orcids":["https://orcid.org/0000-0002-4780-2794"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Guo, Shuyue","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5147825597","display_name":"Tianyu Zheng","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zheng, Tianyu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5148489298","display_name":"Xin Zhou","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhou, Xin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101308310","display_name":"Xingwei Qu","orcid":"https://orcid.org/0000-0002-3265-7133","observed_orcids":["https://orcid.org/0000-0002-3265-7133"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Qu, Xingwei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5074508950","display_name":"Wangchunshu Zhou","orcid":"https://orcid.org/0000-0003-4668-3348","observed_orcids":["https://orcid.org/0000-0003-4668-3348"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhou, Wangchunshu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5148485987","display_name":"Zheng Zhang","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Zheng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100652844","display_name":"Wei Xue","orcid":"https://orcid.org/0000-0002-4942-7748","observed_orcids":["https://orcid.org/0000-0002-4942-7748"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shen, Wei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100652844","display_name":"Wei Xue","orcid":"https://orcid.org/0000-0002-4942-7748","observed_orcids":["https://orcid.org/0000-0002-4942-7748"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xue, Wei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5024599321","display_name":"Chenghua Lin","orcid":"https://orcid.org/0000-0003-3454-2468","observed_orcids":["https://orcid.org/0000-0003-3454-2468"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Qian","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5024599321","display_name":"Chenghua Lin","orcid":"https://orcid.org/0000-0003-3454-2468","observed_orcids":["https://orcid.org/0000-0003-3454-2468"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lin, Chenghua; id_orcid 0000-0003-3454-2468","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5148508931","display_name":"Jian Yang","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yang, Jian","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5148537002","display_name":"Ge Zhang","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Ge","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5148481708","display_name":"Wenhao Huang","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Huang, Wenhao","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":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.71176041,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.4016000032424927,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.4016000032424927,"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/T10028","display_name":"Topic Modeling","score":0.2915000021457672,"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":0.15700000524520874,"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/large-language-model-alignment","display_name":"large language model alignment","score":0.6489999890327454},{"id":"https://openalex.org/keywords/reinforcement-learning-from-human-feedback","display_name":"reinforcement learning from human feedback","score":0.8619999885559082},{"id":"https://openalex.org/keywords/local-uncertainty","display_name":"local uncertainty","score":0.6309999823570251},{"id":"https://openalex.org/keywords/advantage-estimation","display_name":"advantage estimation","score":0.6740000247955322},{"id":"https://openalex.org/keywords/proximal-policy-optimization","display_name":"proximal policy optimization","score":0.6859999895095825},{"id":"https://openalex.org/keywords/inference-efficiency","display_name":"inference efficiency","score":0.7630000114440918},{"id":"https://openalex.org/keywords/reasoning-benchmarks","display_name":"reasoning benchmarks","score":0.4569999873638153}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7444000244140625},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.6557999849319458},{"id":"https://openalex.org/C140779682","wikidata":"https://www.wikidata.org/wiki/Q210868","display_name":"Sampling (signal processing)","level":3,"score":0.5558000206947327},{"id":"https://openalex.org/C113174947","wikidata":"https://www.wikidata.org/wiki/Q2859736","display_name":"Tree (set theory)","level":2,"score":0.5024999976158142},{"id":"https://openalex.org/C108010975","wikidata":"https://www.wikidata.org/wiki/Q500094","display_name":"Pruning","level":2,"score":0.4819999933242798},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.4781000018119812},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4754999876022339},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.45489999651908875},{"id":"https://openalex.org/C2777735758","wikidata":"https://www.wikidata.org/wiki/Q817765","display_name":"Path (computing)","level":2,"score":0.4530999958515167},{"id":"https://openalex.org/C111335779","wikidata":"https://www.wikidata.org/wiki/Q3454686","display_name":"Reduction (mathematics)","level":2,"score":0.43140000104904175},{"id":"https://openalex.org/C115537543","wikidata":"https://www.wikidata.org/wiki/Q165596","display_name":"Cache","level":2,"score":0.4198000133037567},{"id":"https://openalex.org/C52740198","wikidata":"https://www.wikidata.org/wiki/Q1539564","display_name":"Importance sampling","level":3,"score":0.4023999869823456},{"id":"https://openalex.org/C51034333","wikidata":"https://www.wikidata.org/wiki/Q4747796","display_name":"Amortizing loan","level":5,"score":0.4000000059604645},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.3993000090122223},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.39750000834465027},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.39259999990463257},{"id":"https://openalex.org/C2778112365","wikidata":"https://www.wikidata.org/wiki/Q3511065","display_name":"Sequence (biology)","level":2,"score":0.3804999887943268},{"id":"https://openalex.org/C73602740","wikidata":"https://www.wikidata.org/wiki/Q7795822","display_name":"Thompson sampling","level":3,"score":0.37229999899864197},{"id":"https://openalex.org/C173801870","wikidata":"https://www.wikidata.org/wiki/Q201413","display_name":"Heuristic","level":2,"score":0.3312999904155731},{"id":"https://openalex.org/C207390915","wikidata":"https://www.wikidata.org/wiki/Q1230525","display_name":"Divergence (linguistics)","level":2,"score":0.328900009393692},{"id":"https://openalex.org/C197323446","wikidata":"https://www.wikidata.org/wiki/Q331222","display_name":"Oversampling","level":3,"score":0.3278999924659729},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.3174999952316284},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.30970001220703125},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.30820000171661377},{"id":"https://openalex.org/C190290938","wikidata":"https://www.wikidata.org/wiki/Q387015","display_name":"Trie","level":3,"score":0.30820000171661377},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.2924000024795532},{"id":"https://openalex.org/C12725497","wikidata":"https://www.wikidata.org/wiki/Q810247","display_name":"Baseline (sea)","level":2,"score":0.2863999903202057},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.262800008058548},{"id":"https://openalex.org/C2776240099","wikidata":"https://www.wikidata.org/wiki/Q327018","display_name":"Interrogation","level":2,"score":0.2587999999523163}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:repository.hkust.edu.hk:1783.1-173863","doi":null,"is_oa":false,"landing_page_url":"http://repository.hkust.edu.hk/ir/Record/1783.1-173863","pdf_url":null,"source":{"id":"https://openalex.org/S4306401796","display_name":"Rare & Special e-Zone (The Hong Kong University of Science and Technology)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I200769079","host_organization_name":"Hong Kong University of Science and Technology","host_organization_lineage":["https://openalex.org/I200769079"],"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":"Conference paper"},{"id":"pmh:oai:pure.atira.dk:openaire_cris_publications/bba4b107-3f29-47b6-bfc0-cdc09c8fea0d","doi":null,"is_oa":false,"landing_page_url":"https://openreview.net/forum?id=rc6LgDMlPW","pdf_url":null,"source":{"id":"https://openalex.org/S4306400662","display_name":"Research Explorer (The University of Manchester)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I28407311","host_organization_name":"University of Manchester","host_organization_lineage":["https://openalex.org/I28407311"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"Li, Y, Gu, Q, Wen, Z, Li, Z, Yuan, R, Xing, T, Guo, S, Zheng, T, Zhou, X, Qu, X, Zhou, W, Zhang, Z, Shen, W, Xue, W, Liu, Q, Lin, C, Yang, J, Zhang, G & Huang, W 2026, 'TreePO : Enhancing Policy Efficacy and Inference Efficiency with Tree Modeling', Paper presented at Forty-Third International Conference on Machine Learning: ICML 2026, Seoul, 6/07/26 - 11/07/26. < https://openreview.net/forum?id=rc6LgDMlPW >","raw_type":"info:eu-repo/semantics/conferenceObject"}],"best_oa_location":null,"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W7077869697","https://openalex.org/W7165135290","https://openalex.org/W7127428438","https://openalex.org/W7131402120","https://openalex.org/W4412945254","https://openalex.org/W7119470779","https://openalex.org/W7125157117","https://openalex.org/W7123315918","https://openalex.org/W7125994394","https://openalex.org/W7164235203"],"abstract_inverted_index":{"Recent":[0],"advancements":[1],"in":[2,14],"aligning":[3],"large":[4],"language":[5],"models":[6],"via":[7],"reinforcement":[8],"learning":[9],"have":[10],"achieved":[11],"remarkable":[12],"gains":[13],"solving":[15],"complex":[16],"reasoning":[17,32,163],"problems,":[18],"but":[19],"at":[20,191,195,229],"the":[21,86,106,141,155,166,178,182,200],"cost":[22],"of":[23,30,55,143,158,169,177,208],"expensive":[24],"on-policy":[25],"rollouts":[26],"and":[27,60,78,113,132,137,145,149,165,193,223],"limited":[28],"exploration":[29,94],"diverse":[31],"paths.":[33],"In":[34],"this":[35],"work,":[36],"we":[37],"introduce":[38],"TreePO,":[39],"involving":[40],"a":[41,50,100,123,161,205,213],"self-guided":[42],"rollout":[43],"algorithm":[44,103],"that":[45,104,128],"views":[46],"sequence":[47],"generation":[48],"as":[49],"tree-structured":[51],"searching":[52],"process.":[53],"Composed":[54],"dynamic":[56,147],"tree":[57],"sampling":[58,102,179,197],"policy":[59,135],"fixed-length":[61],"segment":[62],"decoding,":[63],"TreePO":[64,83,159,211],"leverages":[65],"local":[66,133],"uncertainty":[67],"to":[68,175,188],"warrant":[69],"additional":[70],"branches.":[71],"By":[72],"amortizing":[73],"computation":[74],"across":[75],"common":[76],"prefixes":[77],"pruning":[79],"low-value":[80],"paths":[81],"early,":[82],"essentially":[84],"reduces":[85],"per-update":[87],"compute":[88,198],"burden":[89,109],"while":[90],"preserving":[91],"or":[92],"enhancing":[93],"diversity.":[95],"Key":[96],"contributions":[97],"include:":[98],"(1)":[99],"segment-wise":[101],"alleviates":[105],"KV":[107],"cache":[108],"through":[110],"contiguous":[111],"segments":[112],"spawns":[114],"new":[115],"branches":[116],"along":[117],"with":[118,220],"an":[119],"early-stop":[120],"mechanism;":[121],"(2)":[122],"tree-based":[124],"segment-level":[125],"advantage":[126],"estimation":[127],"considers":[129],"both":[130],"global":[131],"proximal":[134],"optimization.":[136],"(3)":[138],"analysis":[139],"on":[140,160],"effectiveness":[142],"probability":[144],"quality-driven":[146],"divergence":[148],"fallback":[150],"strategy.":[151],"We":[152],"empirically":[153],"validate":[154],"performance":[156],"gain":[157],"set":[162],"benchmarks":[164],"efficiency":[167],"saving":[168],"GPU":[170],"hours":[171],"from":[172],"22\\%":[173],"up":[174,187],"43\\%":[176],"design":[180],"for":[181,199],"trained":[183],"models,":[184],"meanwhile":[185],"showing":[186],"40\\%":[189],"reduction":[190],"trajectory-level":[192],"35\\%":[194],"token-level":[196],"existing":[201],"models.":[202],"While":[203],"offering":[204],"free":[206],"lunch":[207],"inference":[209],"efficiency,":[210],"reveals":[212],"practical":[214],"path":[215],"toward":[216],"scaling":[217],"RL-based":[218],"post-training":[219],"fewer":[221],"samples":[222],"less":[224],"compute.":[225],"Home":[226],"page":[227],"locates":[228],"https://m-a-p.ai/TreePO.":[230]},"counts_by_year":[],"updated_date":"2026-10-16T09:59:22.235173","created_date":"2026-08-23T00:00:00"}