{"id":"https://openalex.org/W7166836695","doi":"https://doi.org/10.18653/v1/2026.acl-long.966","title":"Beyond the Context Window: Scaling Agentic RL via End-to-end Optimized Context Compression","display_name":"Beyond the Context Window: Scaling Agentic RL via End-to-end Optimized Context Compression","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7166836695","doi":"https://doi.org/10.18653/v1/2026.acl-long.966"},"language":null,"primary_location":{"id":"doi:10.18653/v1/2026.acl-long.966","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.acl-long.966","pdf_url":"https://aclanthology.org/2026.acl-long.966.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2026.acl-long.966.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5139823220","display_name":"Miao Lu","orcid":null},"institutions":[{"id":"https://openalex.org/I97018004","display_name":"Stanford University","ror":"https://ror.org/00f54p054","country_code":"US","type":"education","lineage":["https://openalex.org/I97018004"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Miao Lu","raw_affiliation_strings":["Stanford University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Stanford University","institution_ids":["https://openalex.org/I97018004"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139847452","display_name":"Weiwei Sun","orcid":null},"institutions":[{"id":"https://openalex.org/I74973139","display_name":"Carnegie Mellon University","ror":"https://ror.org/05x2bcf33","country_code":"US","type":"education","lineage":["https://openalex.org/I74973139"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Weiwei Sun","raw_affiliation_strings":["Carnegie Mellon University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Carnegie Mellon University","institution_ids":["https://openalex.org/I74973139"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139803367","display_name":"Weihua Du","orcid":null},"institutions":[{"id":"https://openalex.org/I74973139","display_name":"Carnegie Mellon University","ror":"https://ror.org/05x2bcf33","country_code":"US","type":"education","lineage":["https://openalex.org/I74973139"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Weihua Du","raw_affiliation_strings":["Carnegie Mellon University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Carnegie Mellon University","institution_ids":["https://openalex.org/I74973139"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139781157","display_name":"Zhan Ling","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhan Ling","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139764589","display_name":"Xuesong Yao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xuesong Yao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139757024","display_name":"Kang Liu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Kang Liu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5139717699","display_name":"Jiecao Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jiecao Chen","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"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":0.84700306,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"21074","last_page":"21125"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10444","display_name":"Context-Aware Activity Recognition Systems","score":0.11940000206232071,"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/T10444","display_name":"Context-Aware Activity Recognition Systems","score":0.11940000206232071,"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/T12290","display_name":"Human Motion and Animation","score":0.08420000225305557,"subfield":{"id":"https://openalex.org/subfields/2207","display_name":"Control and Systems Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T12031","display_name":"Speech and dialogue systems","score":0.05429999902844429,"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/context","display_name":"Context (archaeology)","score":0.6431999802589417},{"id":"https://openalex.org/keywords/compression","display_name":"Compression (physics)","score":0.5097000002861023},{"id":"https://openalex.org/keywords/data-compression","display_name":"Data compression","score":0.388700008392334},{"id":"https://openalex.org/keywords/scaling","display_name":"Scaling","score":0.35030001401901245},{"id":"https://openalex.org/keywords/representation","display_name":"Representation (politics)","score":0.2662999927997589}],"concepts":[{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.6431999802589417},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5831999778747559},{"id":"https://openalex.org/C180016635","wikidata":"https://www.wikidata.org/wiki/Q2712821","display_name":"Compression (physics)","level":2,"score":0.5097000002861023},{"id":"https://openalex.org/C78548338","wikidata":"https://www.wikidata.org/wiki/Q2493","display_name":"Data compression","level":2,"score":0.388700008392334},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3571000099182129},{"id":"https://openalex.org/C99844830","wikidata":"https://www.wikidata.org/wiki/Q102441924","display_name":"Scaling","level":2,"score":0.35030001401901245},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.28040000796318054},{"id":"https://openalex.org/C2776359362","wikidata":"https://www.wikidata.org/wiki/Q2145286","display_name":"Representation (politics)","level":3,"score":0.2662999927997589},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.265500009059906},{"id":"https://openalex.org/C183322885","wikidata":"https://www.wikidata.org/wiki/Q17007702","display_name":"Context model","level":3,"score":0.2522999942302704}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2026.acl-long.966","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.acl-long.966","pdf_url":"https://aclanthology.org/2026.acl-long.966.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2026.acl-long.966","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.acl-long.966","pdf_url":"https://aclanthology.org/2026.acl-long.966.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7166836695.pdf","grobid_xml":"https://content.openalex.org/works/W7166836695.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"We":[0],"study":[1],"reinforcement":[2],"learning":[3],"(RL)":[4],"finetuning":[5],"of":[6,185,189],"large":[7],"language":[8],"model":[9],"(LLM)":[10],"agents":[11],"for":[12,169],"long-horizon":[13,131],"multi-turn":[14,25],"tool":[15,60],"use,":[16],"where":[17],"context":[18,40,51,74,84,136,161],"length":[19,162],"quickly":[20],"becomes":[21],"a":[22,72,91,134],"fundamental":[23],"bottleneck.Existing":[24],"RL":[26,100,127],"pipelines":[27],"suffer":[28],"from":[29],"degraded":[30],"instruction":[31],"following,":[32],"excessive":[33],"rollout":[34],"costs,":[35],"and":[36,142],"most":[37],"importantly,":[38],"strict":[39],"limits.In":[41],"this":[42,87,117],"work,":[43],"to":[44,53,70,79,102,164],"address":[45],"these":[46],"challenges,":[47],"we":[48,89],"introduce":[49],"summarization-based":[50],"management":[52],"training.In":[54],"specific,":[55],"it":[56],"periodically":[57],"compresses":[58],"the":[59,77,82,150,155,177],"using":[61],"history":[62],"by":[63],"LLM-generated":[64],"summaries":[65],"that":[66,95,129,146,168,188],"retain":[67],"taskrelevant":[68],"information":[69],"keep":[71],"compact":[73],"while":[75,153],"enabling":[76],"agent":[78],"scale":[80],"beyond":[81,133,187],"fixed":[83,135],"window.Building":[85],"on":[86,138],"formulation,":[88],"derive":[90],"policy":[92],"gradient":[93],"representation":[94],"seamlessly":[96],"enables":[97,130],"standard":[98],"LLM":[99,126],"infrastructures":[101],"optimize":[103],"both":[104],"tool-use":[105],"behaviors":[106],"as":[107,109],"well":[108],"summarization":[110,186],"strategies":[111],"in":[112],"an":[113,125],"end-to-end":[114],"fashion.We":[115],"instantiate":[116],"framework":[118],"with":[119],"SUmmarization":[120],"augmented":[121],"Policy":[122],"Optimization":[123],"(SUPO),":[124],"algorithm":[128],"training":[132,190],"limit.Experiments":[137],"interactive":[139],"function":[140],"calling":[141],"searching":[143,171],"tasks":[144,172],"demonstrate":[145,167],"SUPO":[147,173],"significantly":[148],"improves":[149],"success":[151],"rate":[152],"maintaining":[154],"same":[156],"or":[157],"even":[158],"lower":[159],"working":[160],"compared":[163],"baselines.We":[165],"also":[166],"complex":[170],"can":[174],"further":[175],"improve":[176],"evaluation":[178],"performance":[179],"when":[180],"scaling":[181],"test-time":[182],"maximum":[183],"round":[184],"time.":[191]},"counts_by_year":[],"updated_date":"2026-08-18T07:49:30.821534","created_date":"2026-07-02T00:00:00"}