{"id":"https://openalex.org/W7166118130","doi":"https://doi.org/10.48550/arxiv.2606.27023","title":"Just how sure are you? Improving Verbalized Uncertainty Calibration in Medical VQA","display_name":"Just how sure are you? Improving Verbalized Uncertainty Calibration in Medical VQA","publication_year":2026,"publication_date":"2026-06-25","ids":{"openalex":"https://openalex.org/W7166118130","doi":"https://doi.org/10.48550/arxiv.2606.27023"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2606.27023","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.27023","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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":"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"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2606.27023","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5139446494","display_name":"Eren Senoglu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Senoglu, Eren","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5061447676","display_name":"Federico Toschi","orcid":"https://orcid.org/0000-0001-5935-2332"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Toschi, Federico","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139406657","display_name":"Nicolo Brunello","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Brunello, Nicolo","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5057059969","display_name":"Andrea Sassella","orcid":"https://orcid.org/0000-0001-9067-8496"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Sassella, Andrea","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5069487278","display_name":"Mark Carman","orcid":"https://orcid.org/0000-0001-6575-9737"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Carman, Mark James","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":0.8291000127792358,"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":0.8291000127792358,"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/T13702","display_name":"Machine Learning in Healthcare","score":0.04879999905824661,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.0414000004529953,"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/calibration","display_name":"Calibration","score":0.5303999781608582},{"id":"https://openalex.org/keywords/function","display_name":"Function (biology)","score":0.42750000953674316},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.4074000120162964},{"id":"https://openalex.org/keywords/image","display_name":"Image (mathematics)","score":0.38429999351501465},{"id":"https://openalex.org/keywords/brier-score","display_name":"Brier score","score":0.38190001249313354},{"id":"https://openalex.org/keywords/contrast","display_name":"Contrast (vision)","score":0.3686999976634979},{"id":"https://openalex.org/keywords/pattern-recognition","display_name":"Pattern recognition (psychology)","score":0.3386000096797943},{"id":"https://openalex.org/keywords/question-answering","display_name":"Question answering","score":0.3379000127315521}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7462999820709229},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.538100004196167},{"id":"https://openalex.org/C165838908","wikidata":"https://www.wikidata.org/wiki/Q736777","display_name":"Calibration","level":2,"score":0.5303999781608582},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4487999975681305},{"id":"https://openalex.org/C14036430","wikidata":"https://www.wikidata.org/wiki/Q3736076","display_name":"Function (biology)","level":2,"score":0.42750000953674316},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.4074000120162964},{"id":"https://openalex.org/C115961682","wikidata":"https://www.wikidata.org/wiki/Q860623","display_name":"Image (mathematics)","level":2,"score":0.38429999351501465},{"id":"https://openalex.org/C35405484","wikidata":"https://www.wikidata.org/wiki/Q4967066","display_name":"Brier score","level":2,"score":0.38190001249313354},{"id":"https://openalex.org/C2776502983","wikidata":"https://www.wikidata.org/wiki/Q690182","display_name":"Contrast (vision)","level":2,"score":0.3686999976634979},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.3386000096797943},{"id":"https://openalex.org/C44291984","wikidata":"https://www.wikidata.org/wiki/Q1074173","display_name":"Question answering","level":2,"score":0.3379000127315521},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.33309999108314514},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.32269999384880066},{"id":"https://openalex.org/C2780226545","wikidata":"https://www.wikidata.org/wiki/Q6888030","display_name":"Modality (human\u2013computer interaction)","level":2,"score":0.3156000077724457},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.3111000061035156},{"id":"https://openalex.org/C207390915","wikidata":"https://www.wikidata.org/wiki/Q1230525","display_name":"Divergence (linguistics)","level":2,"score":0.3100000023841858},{"id":"https://openalex.org/C188198153","wikidata":"https://www.wikidata.org/wiki/Q1613840","display_name":"Limiting","level":2,"score":0.30239999294281006},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.29010000824928284},{"id":"https://openalex.org/C160633673","wikidata":"https://www.wikidata.org/wiki/Q355198","display_name":"Pixel","level":2,"score":0.2863999903202057},{"id":"https://openalex.org/C168167062","wikidata":"https://www.wikidata.org/wiki/Q1117970","display_name":"Component (thermodynamics)","level":2,"score":0.27480000257492065},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.27320000529289246},{"id":"https://openalex.org/C57273362","wikidata":"https://www.wikidata.org/wiki/Q576722","display_name":"Decoding methods","level":2,"score":0.2718999981880188}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2606.27023","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.27023","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2606.27023","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.27023","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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":"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"},"sustainable_development_goals":[{"display_name":"Reduced inequalities","id":"https://metadata.un.org/sdg/10","score":0.5917365550994873}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Multimodal":[0],"large":[1],"language":[2,123],"models":[3],"(MLLMs)":[4],"applied":[5],"to":[6,13,54,134],"Medical":[7,146],"Visual":[8],"Question":[9],"Answering":[10],"(VQA)":[11],"tend":[12],"produce":[14],"overconfident":[15],"outputs":[16],"regardless":[17],"of":[18,40,115,139,202],"actual":[19],"correctness,":[20],"and":[21,85,149,155,169,191,195],"existing":[22],"verbalized":[23],"confidence":[24,74],"calibration":[25,57,67,163],"methods,":[26],"developed":[27],"primarily":[28],"for":[29,36,209,215],"text":[30,82,110],"only":[31],"LLMs,":[32],"do":[33],"not":[34],"account":[35],"the":[37,113,116,136,140,184,203,211,216],"multimodal":[38],"nature":[39],"medical":[41],"image":[42,81,107],"understanding.":[43],"This":[44],"work":[45],"proposes":[46],"a":[47,59,64,79,86,98,126],"training":[48,192],"based":[49,88,193],"framework":[50],"that":[51,72,105,199],"finetunes":[52],"MLLMs":[53],"improve":[55],"their":[56],"using":[58],"composite":[60],"loss":[61,204],"function":[62,205],"combining":[63],"Brier":[65],"style":[66],"term,":[68,84],"an":[69],"anchor":[70],"regularizer":[71,131],"prevents":[73],"collapse":[75],"toward":[76],"extreme":[77],"values,":[78],"contrastive":[80],"alignment":[83,93],"KL":[87,129],"model":[89,117,141],"stabilization":[90],"term.":[91],"The":[92],"signal":[94],"is":[95,132,206,218],"derived":[96],"from":[97],"$2":[99],"\\times":[100],"2$":[101],"factorial":[102],"perturbation":[103],"design":[104],"crosses":[106],"presence":[108],"with":[109],"integrity,":[111],"probing":[112],"reliance":[114],"on":[118],"visual":[119],"modality":[120],"input":[121],"versus":[122],"priors.":[124],"Finally,":[125],"top":[127],"K":[128],"divergence":[130],"used":[133],"protect":[135],"answering":[137],"ability":[138],"during":[142],"finetuning.":[143],"Across":[144],"three":[145],"VQA":[147],"benchmarks":[148],"two":[150],"architectures":[151],"(MedGemma":[152],"4B":[153],"IT":[154],"Qwen2":[156],"VL":[157],"7B":[158],"Instruct),":[159],"our":[160],"method":[161],"reduces":[162],"error":[164],"by":[165,172],"60%":[166],"or":[167,174],"more,":[168,175],"improves":[170],"discrimination":[171],"26%":[173],"while":[176],"preserving":[177],"predictive":[178],"accuracy.":[179],"On":[180],"average":[181],"across":[182],"benchmarks,":[183],"technique":[185],"outperforms":[186],"prompting":[187],"based,":[188,190],"sampling":[189],"approaches,":[194],"ablation":[196],"experiments":[197,217],"confirm":[198],"each":[200],"component":[201],"indeed":[207],"necessary":[208],"improving":[210],"calibration.":[212],"All":[213],"code":[214],"publicly":[219],"available.":[220]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-06-27T00:00:00"}