{"id":"https://openalex.org/W7128736715","doi":"https://doi.org/10.48448/6fdx-ne18","title":"Toward Better EHR Reasoning in LLMs: Reinforcement Learning with Expert Attention Guidance","display_name":"Toward Better EHR Reasoning in LLMs: Reinforcement Learning with Expert Attention Guidance","publication_year":2026,"publication_date":"2026-01-07","ids":{"openalex":"https://openalex.org/W7128736715","doi":"https://doi.org/10.48448/6fdx-ne18"},"language":"en","primary_location":{"id":"doi:10.48448/6fdx-ne18","doi":"https://doi.org/10.48448/6fdx-ne18","is_oa":true,"landing_page_url":"https://doi.org/10.48448/6fdx-ne18","pdf_url":null,"source":{"id":"https://openalex.org/S7407053148","display_name":"Underline Science Inc.","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"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":"Audiovisual"},"volume":null,"type":"other","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48448/6fdx-ne18","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5125741655","display_name":"Association for Artificial Intelligence 2026","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Association for Artificial Intelligence 2026","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137994304","display_name":"Hongxin Ding","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ding, Hongxin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5125718284","display_name":"Yue Fang","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Fang, Yue","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5125705276","display_name":"Jiaran Gao","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Gao, Jiaran","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5125690350","display_name":"Yuxin Guo","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Guo, Yuxin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5148862840","display_name":"Xinke Jiang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I20231570","display_name":"Peking University","ror":"https://ror.org/02v51f717","country_code":"CN","type":"education","lineage":["https://openalex.org/I20231570"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jiang, Xinke","raw_affiliation_strings":["Peking University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Peking University","institution_ids":["https://openalex.org/I20231570"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5125687690","display_name":"Weibin Liao","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liao, Weibin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5125691509","display_name":"Liantao Ma","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ma, Liantao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139829275","display_name":"Yasha Wang","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Yasha","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5125690807","display_name":"Yongxin Xu","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I20231570","display_name":"Peking University","ror":"https://ror.org/02v51f717","country_code":"CN","type":"education","lineage":["https://openalex.org/I20231570"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xu, Yongxin","raw_affiliation_strings":["School Of Computer Science, Peking University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School Of Computer Science, Peking University","institution_ids":["https://openalex.org/I20231570"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5124872810","display_name":"Zhibang Yang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I20231570","display_name":"Peking University","ror":"https://ror.org/02v51f717","country_code":"CN","type":"education","lineage":["https://openalex.org/I20231570"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yang, Zhibang","raw_affiliation_strings":["Peking University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Peking University","institution_ids":["https://openalex.org/I20231570"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5121647741","display_name":"Junfeng Zhao","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhao, Junfeng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5125692736","display_name":"Yinghao Zhu","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhu, Yinghao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"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/T13702","display_name":"Machine Learning in Healthcare","score":0.9990000128746033,"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/T13702","display_name":"Machine Learning in Healthcare","score":0.9990000128746033,"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/T11636","display_name":"Artificial Intelligence in Healthcare and Education","score":0.0006000000284984708,"subfield":{"id":"https://openalex.org/subfields/2718","display_name":"Health Informatics"},"field":{"id":"https://openalex.org/fields/27","display_name":"Medicine"},"domain":{"id":"https://openalex.org/domains/4","display_name":"Health Sciences"}},{"id":"https://openalex.org/T10350","display_name":"Electronic Health Records Systems","score":9.999999747378752e-05,"subfield":{"id":"https://openalex.org/subfields/3605","display_name":"Health Information Management"},"field":{"id":"https://openalex.org/fields/36","display_name":"Health Professions"},"domain":{"id":"https://openalex.org/domains/4","display_name":"Health Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/large-language-models","display_name":"large language models","score":0.9900000095367432},{"id":"https://openalex.org/keywords/reinforcement-learning","display_name":"reinforcement learning","score":0.8709999918937683},{"id":"https://openalex.org/keywords/monte-carlo-tree-search","display_name":"Monte Carlo tree search","score":0.7480000257492065},{"id":"https://openalex.org/keywords/two-stage-training","display_name":"two-stage training","score":0.6489999890327454},{"id":"https://openalex.org/keywords/clinical-prediction","display_name":"clinical prediction","score":0.8360000252723694},{"id":"https://openalex.org/keywords/attention-alignment","display_name":"attention alignment","score":0.5820000171661377},{"id":"https://openalex.org/keywords/domain-generalization","display_name":"domain generalization","score":0.621999979019165},{"id":"https://openalex.org/keywords/robustness-to-disturbances","display_name":"robustness to disturbances","score":0.7450000047683716}],"concepts":[{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.67330002784729},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.66839998960495},{"id":"https://openalex.org/C177148314","wikidata":"https://www.wikidata.org/wiki/Q170084","display_name":"Generalization","level":2,"score":0.5547000169754028},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5444999933242798},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.4758000075817108},{"id":"https://openalex.org/C2780719617","wikidata":"https://www.wikidata.org/wiki/Q1030752","display_name":"Salient","level":2,"score":0.4708999991416931},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.4636000096797943},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.41999998688697815},{"id":"https://openalex.org/C2778827112","wikidata":"https://www.wikidata.org/wiki/Q22245680","display_name":"Feature engineering","level":3,"score":0.41850000619888306},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.36970001459121704},{"id":"https://openalex.org/C169258074","wikidata":"https://www.wikidata.org/wiki/Q245748","display_name":"Random forest","level":2,"score":0.3474999964237213},{"id":"https://openalex.org/C113174947","wikidata":"https://www.wikidata.org/wiki/Q2859736","display_name":"Tree (set theory)","level":2,"score":0.33570000529289246},{"id":"https://openalex.org/C105339364","wikidata":"https://www.wikidata.org/wiki/Q2297740","display_name":"Software deployment","level":2,"score":0.33169999718666077},{"id":"https://openalex.org/C2522767166","wikidata":"https://www.wikidata.org/wiki/Q2374463","display_name":"Data science","level":1,"score":0.3133000135421753},{"id":"https://openalex.org/C58328972","wikidata":"https://www.wikidata.org/wiki/Q184609","display_name":"Expert system","level":2,"score":0.31119999289512634},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.3034999966621399},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.28940001130104065},{"id":"https://openalex.org/C3019952477","wikidata":"https://www.wikidata.org/wiki/Q1324077","display_name":"Health records","level":3,"score":0.2825999855995178},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.260699987411499},{"id":"https://openalex.org/C97364631","wikidata":"https://www.wikidata.org/wiki/Q484284","display_name":"Deductive reasoning","level":2,"score":0.25949999690055847},{"id":"https://openalex.org/C183521366","wikidata":"https://www.wikidata.org/wiki/Q7256422","display_name":"Psychology of reasoning","level":4,"score":0.25619998574256897},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.2502000033855438}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48448/6fdx-ne18","doi":"https://doi.org/10.48448/6fdx-ne18","is_oa":true,"landing_page_url":"https://doi.org/10.48448/6fdx-ne18","pdf_url":null,"source":{"id":"https://openalex.org/S7407053148","display_name":"Underline Science Inc.","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"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":"Audiovisual"}],"best_oa_location":{"id":"doi:10.48448/6fdx-ne18","doi":"https://doi.org/10.48448/6fdx-ne18","is_oa":true,"landing_page_url":"https://doi.org/10.48448/6fdx-ne18","pdf_url":null,"source":{"id":"https://openalex.org/S7407053148","display_name":"Underline Science Inc.","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"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":"Audiovisual"},"sustainable_development_goals":[{"score":0.5683000087738037,"id":"https://metadata.un.org/sdg/3","display_name":"Good health and well-being"}],"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/W7128734722","https://openalex.org/W4414856266","https://openalex.org/W7163357454","https://openalex.org/W7102546300","https://openalex.org/W7166807413","https://openalex.org/W4413269536","https://openalex.org/W7125835548","https://openalex.org/W4415319839","https://openalex.org/W7126158642","https://openalex.org/W7084027160"],"abstract_inverted_index":{"Improving":[0],"large":[1],"language":[2],"models":[3,62,107,112],"(LLMs)":[4],"for":[5,13,204],"electronic":[6],"health":[7],"record":[8],"(EHR)":[9],"reasoning":[10,71,98,123,174],"is":[11],"essential":[12],"enabling":[14],"accurate":[15],"and":[16,73,190],"generalizable":[17],"clinical":[18,194,208],"predictions.":[19],"While":[20],"LLMs":[21,50,177],"excel":[22],"at":[23],"medical":[24],"text":[25],"understanding,":[26],"they":[27],"underperform":[28],"on":[29,46,114,162],"EHR-based":[30],"prediction":[31,209],"tasks":[32],"due":[33],"to":[34,66,93,109,131,187,192],"challenges":[35],"in":[36,207],"modeling":[37],"temporally":[38],"structured,":[39],"high-dimensional":[40],"data.":[41,116],"Existing":[42],"approaches":[43],"often":[44],"rely":[45],"hybrid":[47],"paradigms,":[48],"where":[49,104],"serve":[51],"merely":[52],"as":[53],"frozen":[54],"prior":[55],"retrievers":[56],"while":[57,183],"downstream":[58],"deep":[59],"learning":[60,145],"(DL)":[61],"handle":[63],"prediction,":[64],"failing":[65],"improve":[67],"the":[68,75,134,141,148,171,199],"LLM\u2019s":[69,135,149],"intrinsic":[70,172],"capacity":[72],"inheriting":[74],"generalization":[76,191],"limitations":[77],"of":[78,176,181,202],"DL":[79,111],"models.":[80,159],"To":[81],"this":[82],"end,":[83],"we":[84],"propose":[85],"EAG-RL,":[86],"a":[87],"novel":[88],"two-stage":[89],"training":[90],"framework":[91],"designed":[92],"intrinsically":[94],"enhance":[95],"LLMs\u2019":[96],"EHR":[97,106,115,158,165,173],"ability":[99,175],"through":[100],"expert":[101,105,157],"attention":[102,150],"guidance,":[103],"refer":[108],"task-specific":[110],"trained":[113],"Concretely,":[117],"EAG-RL":[118,138,169,203],"first":[119],"constructs":[120],"high-quality,":[121],"stepwise":[122],"trajectories":[124],"using":[125],"expert-guided":[126],"Monte":[127],"Carlo":[128],"Tree":[129],"Search":[130],"effectively":[132],"initialize":[133],"policy.":[136],"Then,":[137],"further":[139],"optimizes":[140],"policy":[142],"via":[143],"reinforcement":[144],"by":[146,156,178],"aligning":[147],"with":[151],"clinically":[152],"salient":[153],"features":[154],"identified":[155],"Extensive":[160],"experiments":[161],"two":[163],"real-world":[164,205],"datasets":[166],"show":[167],"that":[168],"improves":[170],"an":[179],"average":[180],"14.62%,":[182],"also":[184],"enhancing":[185],"robustness":[186],"feature":[188],"perturbations":[189],"unseen":[193],"domains.":[195],"These":[196],"results":[197],"demonstrate":[198],"practical":[200],"potential":[201],"deployment":[206],"tasks.":[210]},"counts_by_year":[],"updated_date":"2026-10-28T10:33:24.094632","created_date":"2026-02-13T00:00:00"}