{"id":"https://openalex.org/W7167641037","doi":"https://doi.org/10.48550/arxiv.2607.04625","title":"Hierarchical Evidence-Driven Reasoning for Long Document Understanding","display_name":"Hierarchical Evidence-Driven Reasoning for Long Document Understanding","publication_year":2026,"publication_date":"2026-07-06","ids":{"openalex":"https://openalex.org/W7167641037","doi":"https://doi.org/10.48550/arxiv.2607.04625"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2607.04625","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2607.04625","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"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":"Preprint"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2607.04625","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5101368175","display_name":"Junyu Xiong","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xiong, Junyu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140170052","display_name":"Yonghui Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Yonghui","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140193462","display_name":"Rongjian Gu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Gu, Rongjian","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140203951","display_name":"Chenyu Liu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Chenyu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140169995","display_name":"Bing Yin","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yin, Bing","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140180673","display_name":"Wengang Zhou","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhou, Wengang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5140183232","display_name":"Houqiang Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Houqiang","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":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":0.9217000007629395,"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":0.9217000007629395,"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/T10627","display_name":"Advanced Image and Video Retrieval Techniques","score":0.027400000020861626,"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":0.010900000110268593,"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":0.5900999903678894},{"id":"https://openalex.org/keywords/semantics","display_name":"Semantics (computer science)","score":0.4934000074863434},{"id":"https://openalex.org/keywords/pipeline","display_name":"Pipeline (software)","score":0.4278999865055084},{"id":"https://openalex.org/keywords/similarity","display_name":"Similarity (geometry)","score":0.4023999869823456},{"id":"https://openalex.org/keywords/baseline","display_name":"Baseline (sea)","score":0.398499995470047},{"id":"https://openalex.org/keywords/ranking","display_name":"Ranking (information retrieval)","score":0.3393000066280365},{"id":"https://openalex.org/keywords/core","display_name":"Core (optical fiber)","score":0.30559998750686646}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8388000130653381},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.5900999903678894},{"id":"https://openalex.org/C23123220","wikidata":"https://www.wikidata.org/wiki/Q816826","display_name":"Information retrieval","level":1,"score":0.5819000005722046},{"id":"https://openalex.org/C184337299","wikidata":"https://www.wikidata.org/wiki/Q1437428","display_name":"Semantics (computer science)","level":2,"score":0.4934000074863434},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4311000108718872},{"id":"https://openalex.org/C43521106","wikidata":"https://www.wikidata.org/wiki/Q2165493","display_name":"Pipeline (software)","level":2,"score":0.4278999865055084},{"id":"https://openalex.org/C103278499","wikidata":"https://www.wikidata.org/wiki/Q254465","display_name":"Similarity (geometry)","level":3,"score":0.4023999869823456},{"id":"https://openalex.org/C12725497","wikidata":"https://www.wikidata.org/wiki/Q810247","display_name":"Baseline (sea)","level":2,"score":0.398499995470047},{"id":"https://openalex.org/C189430467","wikidata":"https://www.wikidata.org/wiki/Q7293293","display_name":"Ranking (information retrieval)","level":2,"score":0.3393000066280365},{"id":"https://openalex.org/C2164484","wikidata":"https://www.wikidata.org/wiki/Q5170150","display_name":"Core (optical fiber)","level":2,"score":0.30559998750686646},{"id":"https://openalex.org/C2779982483","wikidata":"https://www.wikidata.org/wiki/Q6094420","display_name":"Iterative refinement","level":2,"score":0.30219998955726624},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.30140000581741333},{"id":"https://openalex.org/C1667742","wikidata":"https://www.wikidata.org/wiki/Q10927554","display_name":"Image retrieval","level":3,"score":0.29660001397132874},{"id":"https://openalex.org/C131918245","wikidata":"https://www.wikidata.org/wiki/Q1409090","display_name":"Fetch","level":2,"score":0.2806999981403351},{"id":"https://openalex.org/C2778112365","wikidata":"https://www.wikidata.org/wiki/Q3511065","display_name":"Sequence (biology)","level":2,"score":0.271699994802475},{"id":"https://openalex.org/C127705205","wikidata":"https://www.wikidata.org/wiki/Q5748245","display_name":"Heuristics","level":2,"score":0.2694999873638153},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.26010000705718994},{"id":"https://openalex.org/C130318100","wikidata":"https://www.wikidata.org/wiki/Q2268914","display_name":"Semantic similarity","level":2,"score":0.2563000023365021}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2607.04625","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2607.04625","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"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":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2607.04625","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2607.04625","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"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":"Preprint"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Retrieval-Augmented":[0],"Generation":[1],"(RAG)":[2],"streamlines":[3],"long-document":[4],"understanding":[5],"by":[6,188],"leveraging":[7,111],"retrieval":[8,54,110],"mechanisms":[9],"to":[10,14,97,115,136,153],"restrict":[11],"input":[12],"images":[13],"a":[15,73,89,112,129],"highly":[16],"curated":[17],"subset.":[18],"However,":[19],"existing":[20,179],"multimodal":[21,76,113],"RAG":[22,77],"pipelines":[23,49],"primarily":[24],"face":[25],"two":[26],"critical":[27],"challenges:":[28],"first,":[29],"standard":[30],"semantic":[31,121],"similarity":[32],"retrievers":[33],"frequently":[34],"fetch":[35,116],"topically":[36],"overlapping":[37],"yet":[38],"answer-void":[39],"distractor":[40],"pages":[41,118],"that":[42,148],"mislead":[43],"downstream":[44],"generation;":[45],"second,":[46],"rigid":[47],"single-pass":[48],"heavily":[50],"depend":[51],"on":[52,120],"initial":[53],"success,":[55],"where":[56],"any":[57],"omission":[58],"of":[59,173,191],"core":[60],"evidence":[61],"inevitably":[62],"causes":[63],"cascading":[64],"errors.":[65],"To":[66],"address":[67],"these":[68],"challenges,":[69],"we":[70],"introduce":[71],"HIEVI-RAG,":[72],"hierarchical,":[74],"evidence-driven":[75],"framework":[78],"for":[79],"closed-domain":[80],"document":[81],"understanding.":[82],"HIEVI-RAG":[83],"systematically":[84],"factorizes":[85],"complex":[86],"queries":[87,101],"into":[88,102],"cooperative":[90],"four-stage":[91],"pipeline:":[92],"(1)":[93],"hierarchical":[94],"question":[95],"decomposition":[96],"break":[98],"multi-hop":[99],"root":[100],"atomic":[103],"child":[104],"questions;":[105],"(2)":[106],"coarse":[107],"visual":[108],"page":[109,125],"retriever":[114],"candidate":[117],"based":[119],"similarity;":[122],"(3)":[123],"fine-grained":[124],"verification":[126],"via":[127],"EVIAGENT,":[128],"specialized":[130],"multi-page":[131],"verifier":[132],"trained":[133],"with":[134],"GRPO":[135],"execute":[137,154],"cross-page":[138],"reasoning":[139,157],"over":[140,158],"multi-image":[141],"blocks;":[142],"and":[143,171,182],"(4)":[144],"memory-guided":[145],"iterative":[146],"generation":[147],"leverages":[149],"accumulated":[150],"sub-question":[151],"context":[152],"multi-round,":[155],"dynamic":[156],"the":[159,168,184],"prioritized":[160],"sequence.":[161],"Extensive":[162],"evaluations":[163],"across":[164],"four":[165],"benchmarks":[166],"demonstrate":[167],"robust":[169],"efficacy":[170],"synergy":[172],"our":[174],"framework,":[175],"which":[176],"significantly":[177],"outperforms":[178],"open-source":[180],"baselines":[181],"exceeds":[183],"strongest":[185],"reported":[186],"baseline":[187],"an":[189],"average":[190],"8.05%":[192],"in":[193],"accuracy.":[194]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2026-07-08T00:00:00"}