{"id":"https://openalex.org/W7151835571","doi":"https://doi.org/10.48550/arxiv.2604.05808","title":"Hierarchical Reinforcement Learning with Augmented Step-Level Transitions for LLM Agents","display_name":"Hierarchical Reinforcement Learning with Augmented Step-Level Transitions for LLM Agents","publication_year":2026,"publication_date":"2026-04-07","ids":{"openalex":"https://openalex.org/W7151835571","doi":"https://doi.org/10.48550/arxiv.2604.05808"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2604.05808","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.05808","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"article"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2604.05808","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5111270851","display_name":"Shuai Zhen","orcid":null},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Zhen, Shuai","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5133233217","display_name":"Yanhua Yu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yu, Yanhua","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130725963","display_name":"Ruopei Guo","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Guo, Ruopei","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5133163478","display_name":"Nan Cheng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Cheng, Nan","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5133222008","display_name":"Yang Deng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Deng, Yang","raw_affiliation_strings":[],"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":5,"corresponding_author_ids":["https://openalex.org/A5111270851"],"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.14830000698566437,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.14830000698566437,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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":0.14169999957084656,"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":0.12470000237226486,"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/reinforcement-learning","display_name":"Reinforcement learning","score":0.805899977684021},{"id":"https://openalex.org/keywords/generalization","display_name":"Generalization","score":0.6462000012397766},{"id":"https://openalex.org/keywords/code","display_name":"Code (set theory)","score":0.5123999714851379},{"id":"https://openalex.org/keywords/security-token","display_name":"Security token","score":0.4311999976634979},{"id":"https://openalex.org/keywords/reinforcement","display_name":"Reinforcement","score":0.30880001187324524}],"concepts":[{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.805899977684021},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7533000111579895},{"id":"https://openalex.org/C177148314","wikidata":"https://www.wikidata.org/wiki/Q170084","display_name":"Generalization","level":2,"score":0.6462000012397766},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5523999929428101},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.5123999714851379},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.4311999976634979},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.37400001287460327},{"id":"https://openalex.org/C67203356","wikidata":"https://www.wikidata.org/wiki/Q1321905","display_name":"Reinforcement","level":2,"score":0.30880001187324524},{"id":"https://openalex.org/C136197465","wikidata":"https://www.wikidata.org/wiki/Q1729295","display_name":"Variety (cybernetics)","level":2,"score":0.2775000035762787},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.2732999920845032},{"id":"https://openalex.org/C43126263","wikidata":"https://www.wikidata.org/wiki/Q128751","display_name":"Source code","level":2,"score":0.25369998812675476}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2604.05808","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.05808","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"article"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2604.05808","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.05808","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"article"},"sustainable_development_goals":[{"id":"https://metadata.un.org/sdg/16","score":0.7729069590568542,"display_name":"Peace, Justice and strong institutions"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Large":[0],"language":[1],"model":[2],"(LLM)":[3],"agents":[4,17],"have":[5],"demonstrated":[6],"strong":[7],"capabilities":[8],"in":[9,26,126],"complex":[10],"interactive":[11],"decision-making":[12],"tasks.":[13],"However,":[14],"existing":[15],"LLM":[16],"typically":[18],"rely":[19],"on":[20,52,114],"increasingly":[21],"long":[22],"interaction":[23,58,86],"histories,":[24],"resulting":[25],"high":[27],"computational":[28],"cost":[29],"and":[30,83,109,116,130],"limited":[31],"scalability.":[32],"In":[33],"this":[34],"paper,":[35],"we":[36],"propose":[37],"STEP-HRL,":[38],"a":[39,76,93],"hierarchical":[40],"reinforcement":[41],"learning":[42,48],"(HRL)":[43],"framework":[44],"that":[45,121],"enables":[46],"step-level":[47,104],"by":[49],"conditioning":[50],"only":[51],"single-step":[53],"transitions":[54,105],"rather":[55],"than":[56],"full":[57],"histories.":[59],"STEP-HRL":[60,122],"structures":[61],"tasks":[62],"hierarchically,":[63],"using":[64],"completed":[65],"subtasks":[66],"to":[67,91],"represent":[68],"global":[69],"progress":[70,78],"of":[71,96,128],"overall":[72],"task.":[73],"By":[74],"introducing":[75],"local":[77,97],"module,":[79],"it":[80],"also":[81],"iteratively":[82],"selectively":[84],"summarizes":[85],"history":[87],"within":[88],"each":[89],"subtask":[90],"produce":[92],"compact":[94],"summary":[95],"progress.":[98],"Together,":[99],"these":[100],"components":[101],"yield":[102],"augmented":[103],"for":[106],"both":[107],"high-level":[108],"low-level":[110],"policies.":[111],"Experimental":[112],"results":[113],"ScienceWorld":[115],"ALFWorld":[117],"benchmarks":[118],"consistently":[119],"demonstrate":[120],"substantially":[123],"outperforms":[124],"baselines":[125],"terms":[127],"performance":[129],"generalization":[131],"while":[132],"reducing":[133],"token":[134],"usage.":[135],"Our":[136],"code":[137],"is":[138],"available":[139],"at":[140],"https://github.com/TonyStark042/STEP-HRL.":[141]},"counts_by_year":[],"updated_date":"2026-04-09T06:13:59.934233","created_date":"2026-04-09T00:00:00"}