{"id":"https://openalex.org/W7207555830","doi":"https://doi.org/10.48550/arxiv.2609.01687","title":"Learning Evidence Sufficiency Boundaries for Selective Answering in Grounded Multi-Hop QA","display_name":"Learning Evidence Sufficiency Boundaries for Selective Answering in Grounded Multi-Hop QA","publication_year":2026,"publication_date":"2026-09-01","ids":{"openalex":"https://openalex.org/W7207555830","doi":"https://doi.org/10.48550/arxiv.2609.01687"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2609.01687","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.01687","pdf_url":"https://arxiv.org/pdf/2609.01687","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,"listed_in":[],"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2609.01687","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5149899049","display_name":"Haruto Sato","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Sato, Haruto","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5149623534","display_name":"Yuki Tanaka","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tanaka, Yuki","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5149135555","display_name":"Ren Nakamura","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Nakamura, Ren","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5149266681","display_name":"Aoi Kobayashi","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Kobayashi, Aoi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5149365355","display_name":"Mei Ito","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ito, Mei","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":true,"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/T10028","display_name":"Topic Modeling","score":0.8999000191688538,"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/T10028","display_name":"Topic Modeling","score":0.8999000191688538,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.07829999923706055,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.006300000008195639,"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/multi-hop-question-answering","display_name":"multi-hop question answering","score":0.9549999833106995},{"id":"https://openalex.org/keywords/abstention","display_name":"abstention","score":0.6019999980926514},{"id":"https://openalex.org/keywords/chain-of-custody","display_name":"chain of custody","score":0.6620000004768372},{"id":"https://openalex.org/keywords/llm-fine-tuning","display_name":"LLM fine-tuning","score":0.4959999918937683}],"concepts":[{"id":"https://openalex.org/C44291984","wikidata":"https://www.wikidata.org/wiki/Q1074173","display_name":"Question answering","level":2,"score":0.6607999801635742},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.6395000219345093},{"id":"https://openalex.org/C62354387","wikidata":"https://www.wikidata.org/wiki/Q875399","display_name":"Boundary (topology)","level":2,"score":0.60589998960495},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5108000040054321},{"id":"https://openalex.org/C100660578","wikidata":"https://www.wikidata.org/wiki/Q18733","display_name":"Recall","level":2,"score":0.4027999937534332},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.39410001039505005},{"id":"https://openalex.org/C132964779","wikidata":"https://www.wikidata.org/wiki/Q2110223","display_name":"Raw data","level":2,"score":0.3903000056743622},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3531999886035919},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.3456999957561493},{"id":"https://openalex.org/C156325361","wikidata":"https://www.wikidata.org/wiki/Q1152864","display_name":"Grounded theory","level":3,"score":0.34279999136924744},{"id":"https://openalex.org/C15744967","wikidata":"https://www.wikidata.org/wiki/Q9418","display_name":"Psychology","level":0,"score":0.3328999876976013},{"id":"https://openalex.org/C133462117","wikidata":"https://www.wikidata.org/wiki/Q4929239","display_name":"Data collection","level":2,"score":0.3181999921798706},{"id":"https://openalex.org/C145420912","wikidata":"https://www.wikidata.org/wiki/Q853077","display_name":"Mathematics education","level":1,"score":0.2768000066280365},{"id":"https://openalex.org/C113336015","wikidata":"https://www.wikidata.org/wiki/Q574010","display_name":"Complete information","level":2,"score":0.2754000127315521},{"id":"https://openalex.org/C2780735816","wikidata":"https://www.wikidata.org/wiki/Q28324931","display_name":"Incremental learning","level":2,"score":0.2653999924659729}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2609.01687","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.01687","pdf_url":"https://arxiv.org/pdf/2609.01687","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,"listed_in":[],"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2609.01687","doi":"https://doi.org/10.48550/arxiv.2609.01687","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2609.01687","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,"listed_in":[],"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2609.01687","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.01687","pdf_url":"https://arxiv.org/pdf/2609.01687","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,"listed_in":[],"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7207555830.pdf","grobid_xml":"https://content.openalex.org/works/W7207555830.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W7214945477","https://openalex.org/W6910694488","https://openalex.org/W7139955634","https://openalex.org/W4287083780","https://openalex.org/W3021762271","https://openalex.org/W7155013355","https://openalex.org/W3178327348","https://openalex.org/W3021358742","https://openalex.org/W7214889621","https://openalex.org/W2972838397"],"abstract_inverted_index":{"Grounded":[0],"question":[1],"answering":[2,34,194],"systems":[3],"should":[4,45,204],"answer":[5,28,53,63,103],"only":[6],"when":[7,54,65,196],"the":[8,12,40,55,62,87,141,146,165,179,199],"supplied":[9],"evidence":[10,23,36,67,83,108,200],"supports":[11],"answer.":[13],"In":[14],"multi-hop":[15],"QA,":[16],"this":[17],"requirement":[18],"is":[19,68],"difficult":[20],"because":[21],"partial":[22],"can":[24],"make":[25],"an":[26],"unsupported":[27,48],"appear":[29],"plausible.":[30],"We":[31,70,106],"study":[32],"selective":[33,193],"through":[35],"sufficiency":[37],"boundaries:":[38],"for":[39,157,178],"same":[41,180],"question,":[42],"a":[43,76,96,158],"model":[44],"abstain":[46],"under":[47],"or":[49],"partially":[50],"supported":[51],"context,":[52],"context":[56],"first":[57],"becomes":[58],"sufficient,":[59],"and":[60,85,102,113,124,133],"keep":[61],"stable":[64],"redundant":[66],"added.":[69],"introduce":[71],"Evidence":[72,136],"Sufficiency":[73,137],"Boundary":[74,138],"Training,":[75],"generation-native":[77],"training":[78,197],"framework":[79],"that":[80,191],"constructs":[81],"ordered":[82],"chains":[84,109],"supervises":[86],"abstain-to-answer":[88],"transition":[89],"directly.":[90],"The":[91,188],"method":[92],"combines":[93],"level":[94,201],"supervision,":[95],"boundary":[97,143],"flip":[98,150],"margin,":[99],"post-boundary":[100],"stability,":[101],"recall":[104],"protection.":[105],"build":[107],"from":[110],"HotpotQA,":[111],"2WikiMultiHopQA,":[112],"MuSiQue,":[114],"then":[115],"evaluate":[116],"models":[117],"with":[118,149,155,176],"chain":[119],"metrics,":[120],"raw":[121,185],"QA":[122,186],"utility,":[123],"unsupported-answer":[125,168],"rates":[126],"on":[127,170],"external":[128,171],"non-answerable":[129,172],"sets.":[130],"With":[131],"Qwen2.5-3B-Instruct":[132],"LoRA":[134],"adaptation,":[135],"Training":[139],"gives":[140],"strongest":[142],"localization":[144],"among":[145],"tested":[147],"systems,":[148],"accuracy":[151],"of":[152],"0.807":[153],"compared":[154,175],"0.781":[156],"token-level":[159],"abstention":[160],"baseline.":[161],"It":[162],"also":[163],"achieves":[164],"lowest":[166],"overall":[167],"rate":[169],"evaluation,":[173],"0.095":[174],"0.101":[177],"baseline,":[181],"while":[182],"retaining":[183],"competitive":[184],"F1.":[187],"results":[189],"show":[190],"grounded":[192],"improves":[195],"marks":[198],"where":[202],"refusal":[203],"give":[205],"way":[206],"to":[207],"answering.":[208]},"counts_by_year":[],"updated_date":"2026-10-28T10:33:24.094632","created_date":"2026-09-04T00:00:00"}