{"id":"https://openalex.org/W7166815917","doi":"https://doi.org/10.18653/v1/2026.findings-acl.856","title":"Context-attended Adversarial Reinforcement Learning for Robust Multi-step Retrieval Augmented Generation","display_name":"Context-attended Adversarial Reinforcement Learning for Robust Multi-step Retrieval Augmented Generation","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7166815917","doi":"https://doi.org/10.18653/v1/2026.findings-acl.856"},"language":null,"primary_location":{"id":"doi:10.18653/v1/2026.findings-acl.856","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.856","pdf_url":"https://aclanthology.org/2026.findings-acl.856.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":"Findings of the Association for Computational Linguistics: ACL 2026","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2026.findings-acl.856.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5139832199","display_name":"Yingtao Ren","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yingtao Ren","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139777408","display_name":"Xiao Luo","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xiao Luo","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139846287","display_name":"Yu-Cheng Chang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yu-Cheng Chang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5139772465","display_name":"Chin-teng Lin","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chin-teng Lin","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":18.8205,"has_fulltext":true,"cited_by_count":1,"citation_normalized_percentile":{"value":0.99138067,"is_in_top_1_percent":true,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":94,"max":97},"biblio":{"volume":null,"issue":null,"first_page":"17305","last_page":"17319"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.4043999910354614,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.4043999910354614,"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/T11448","display_name":"Face recognition and analysis","score":0.09430000185966492,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.09189999848604202,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/adversarial-system","display_name":"Adversarial system","score":0.5504000186920166},{"id":"https://openalex.org/keywords/reinforcement-learning","display_name":"Reinforcement learning","score":0.5152999758720398},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.3732999861240387},{"id":"https://openalex.org/keywords/matching","display_name":"Matching (statistics)","score":0.34950000047683716},{"id":"https://openalex.org/keywords/action","display_name":"Action (physics)","score":0.33629998564720154},{"id":"https://openalex.org/keywords/robustness","display_name":"Robustness (evolution)","score":0.27570000290870667}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6473000049591064},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.6003999710083008},{"id":"https://openalex.org/C37736160","wikidata":"https://www.wikidata.org/wiki/Q1801315","display_name":"Adversarial system","level":2,"score":0.5504000186920166},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.5152999758720398},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.3732999861240387},{"id":"https://openalex.org/C165064840","wikidata":"https://www.wikidata.org/wiki/Q1321061","display_name":"Matching (statistics)","level":2,"score":0.34950000047683716},{"id":"https://openalex.org/C2780791683","wikidata":"https://www.wikidata.org/wiki/Q846785","display_name":"Action (physics)","level":2,"score":0.33629998564720154},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.32330000400543213},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.27570000290870667},{"id":"https://openalex.org/C2775924081","wikidata":"https://www.wikidata.org/wiki/Q55608371","display_name":"Control (management)","level":2,"score":0.2685000002384186},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.26460000872612},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.2599000036716461}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2026.findings-acl.856","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.856","pdf_url":"https://aclanthology.org/2026.findings-acl.856.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":"Findings of the Association for Computational Linguistics: ACL 2026","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2026.findings-acl.856","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.856","pdf_url":"https://aclanthology.org/2026.findings-acl.856.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":"Findings of the Association for Computational Linguistics: ACL 2026","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G3814827824","display_name":"Fluent Mobility for the Blind Individual Using Multimodal Auditory Sensory Augmentation","funder_award_id":"2021183","funder_id":"https://openalex.org/F4320334705","funder_display_name":"National Health and Medical Research Council"}],"funders":[{"id":"https://openalex.org/F4320334626","display_name":"Medical Research Council","ror":"https://ror.org/03x94j517"},{"id":"https://openalex.org/F4320334705","display_name":"National Health and Medical Research Council","ror":"https://ror.org/011kf5r70"}],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7166815917.pdf","grobid_xml":"https://content.openalex.org/works/W7166815917.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Multi-step":[0],"retrieval-augmented":[1],"generation":[2,128],"has":[3],"attracted":[4],"increasing":[5],"attention":[6],"due":[7,66],"to":[8,11,31,50,61,67,96,117,144,155,158,173],"its":[9],"capacity":[10],"improve":[12],"the":[13,23,52,68,119,127,146,160,163,175],"factuality":[14],"of":[15,25,70,92,121,129,162,177],"large":[16],"language":[17],"models":[18],"with":[19,107,132],"iterative":[20],"retrieved":[21],"knowledge.However,":[22],"performance":[24,55],"multi-step":[26,86,167],"RAG":[27,87,168],"systems":[28,88],"is":[29,95,124,150,186],"susceptible":[30],"potential":[32],"retrieval":[33],"noise":[34],"and":[35],"fabricated":[36],"documents":[37],"in":[38,181],"real-world":[39],"scenarios.Current":[40],"approaches":[41],"usually":[42],"utilize":[43],"supervised":[44],"fine-tuning":[45],"on":[46,100,166],"predetermined":[47],"noisy":[48,136,183],"contexts":[49],"enhance":[51],"robustness.However,":[53],"their":[54],"remains":[56],"inadequate":[57],"when":[58],"it":[59],"comes":[60],"more":[62],"complicated":[63],"long-context":[64],"scenarios":[65],"lack":[69],"adaptability.Towards":[71],"this":[72],"end,":[73],"we":[74],"propose":[75],"a":[76,114],"novel":[77],"framework":[78,149],"named":[79],"Context-attended":[80],"Adversarial":[81],"Reinforcement":[82],"Learning":[83],"(CARE)":[84],"for":[85,126,141],"against":[89],"attacks.The":[90],"core":[91],"our":[93,111,178],"CARE":[94,112,180],"conduct":[97],"reinforcement":[98,142],"learning":[99,143],"adversarial":[101,130],"samples":[102,131,137,157],"which":[103,123],"are":[104,138,171],"alternatingly":[105,152],"enhanced":[106],"text":[108,133],"gradients.In":[109],"particular,":[110],"includes":[113],"reward":[115],"model":[116],"identify":[118],"accuracy":[120],"responses,":[122],"minimized":[125],"gradients.These":[134],"context-attended":[135],"then":[139],"utilized":[140],"maximize":[145],"rewards.The":[147],"whole":[148],"conducted":[151,172],"from":[153],"easy":[154],"hard":[156],"ensure":[159],"smoothness":[161],"optimization.Extensive":[164],"experiments":[165],"benchmark":[169],"datasets":[170],"validate":[174],"superiority":[176],"proposed":[179],"multiple":[182],"scenarios.Our":[184],"code":[185],"available":[187],"at":[188],"https://github.com/yingtaoren/CARE.":[189]},"counts_by_year":[{"year":2026,"cited_by_count":1}],"updated_date":"2026-08-08T01:25:22.217667","created_date":"2026-07-02T00:00:00"}