{"id":"https://openalex.org/W7214780475","doi":"https://doi.org/10.48550/arxiv.2609.32185","title":"Contamination, Prior, or Evidence? Decomposing and Training Evidence Use in Whole-Slide Vision-Language Models","display_name":"Contamination, Prior, or Evidence? Decomposing and Training Evidence Use in Whole-Slide Vision-Language Models","publication_year":2026,"publication_date":"2026-09-26","ids":{"openalex":"https://openalex.org/W7214780475","doi":"https://doi.org/10.48550/arxiv.2609.32185"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2609.32185","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.32185","pdf_url":"https://arxiv.org/pdf/2609.32185","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/2609.32185","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5153440534","display_name":"Wenhao Zhang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I51556381","display_name":"University of Virginia","ror":"https://ror.org/0153tk833","country_code":"US","type":"education","lineage":["https://openalex.org/I51556381"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Zhang, Wenhao","raw_affiliation_strings":["University of Virginia , Charlottesville"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Virginia , Charlottesville","institution_ids":["https://openalex.org/I51556381"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140786739","display_name":"Zhongliang Zhou","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I1285764155","display_name":"Merck & Co., Inc., Rahway, NJ, USA (United States)","ror":"https://ror.org/02891sr49","country_code":"US","type":"company","lineage":["https://openalex.org/I1285764155"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Zhou, Zhongliang","raw_affiliation_strings":["Merck & Co. , Inc"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Merck & Co. , Inc","institution_ids":["https://openalex.org/I1285764155"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5153517301","display_name":"Shiyuan Zhang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I51556381","display_name":"University of Virginia","ror":"https://ror.org/0153tk833","country_code":"US","type":"education","lineage":["https://openalex.org/I51556381"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Zhang, Shiyuan","raw_affiliation_strings":["University of Virginia , Charlottesville"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Virginia , Charlottesville","institution_ids":["https://openalex.org/I51556381"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5153540308","display_name":"Yiqing Yang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I51556381","display_name":"University of Virginia","ror":"https://ror.org/0153tk833","country_code":"US","type":"education","lineage":["https://openalex.org/I51556381"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Yang, Yiqing","raw_affiliation_strings":["University of Virginia , Charlottesville"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Virginia , Charlottesville","institution_ids":["https://openalex.org/I51556381"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5064569076","display_name":"P Wang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I51556381","display_name":"University of Virginia","ror":"https://ror.org/0153tk833","country_code":"US","type":"education","lineage":["https://openalex.org/I51556381"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Wang, Pinqiao","raw_affiliation_strings":["University of Virginia , Charlottesville"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Virginia , Charlottesville","institution_ids":["https://openalex.org/I51556381"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5001312304","display_name":"Lehan Yang","orcid":"https://orcid.org/0009-0003-7915-2014","observed_orcids":["https://orcid.org/0009-0003-7915-2014","https://orcid.org/0009-0000-0137-685X"]},"institutions":[{"id":"https://openalex.org/I51556381","display_name":"University of Virginia","ror":"https://ror.org/0153tk833","country_code":"US","type":"education","lineage":["https://openalex.org/I51556381"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Yang, Lehan","raw_affiliation_strings":["University of Virginia , Charlottesville"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Virginia , Charlottesville","institution_ids":["https://openalex.org/I51556381"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5013719827","display_name":"Hanyin Wang","orcid":"https://orcid.org/0000-0001-9884-9683","observed_orcids":["https://orcid.org/0000-0001-9884-9683"]},"institutions":[{"id":"https://openalex.org/I1285764155","display_name":"Merck & Co., Inc., Rahway, NJ, USA (United States)","ror":"https://ror.org/02891sr49","country_code":"US","type":"company","lineage":["https://openalex.org/I1285764155"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Wang, Hanyin","raw_affiliation_strings":["Merck & Co. , Inc"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Merck & Co. , Inc","institution_ids":["https://openalex.org/I1285764155"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136553521","display_name":"John Kang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I1285764155","display_name":"Merck & Co., Inc., Rahway, NJ, USA (United States)","ror":"https://ror.org/02891sr49","country_code":"US","type":"company","lineage":["https://openalex.org/I1285764155"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Kang, John","raw_affiliation_strings":["Merck & Co. , Inc"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Merck & Co. , Inc","institution_ids":["https://openalex.org/I1285764155"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5153453918","display_name":"Sheng Li","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I51556381","display_name":"University of Virginia","ror":"https://ror.org/0153tk833","country_code":"US","type":"education","lineage":["https://openalex.org/I51556381"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Li, Sheng","raw_affiliation_strings":["University of Virginia , Charlottesville"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Virginia , Charlottesville","institution_ids":["https://openalex.org/I51556381"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"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/T10862","display_name":"AI in cancer detection","score":0.9916999936103821,"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/T10862","display_name":"AI in cancer detection","score":0.9916999936103821,"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":0.0052999998442828655,"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/T11636","display_name":"Artificial Intelligence in Healthcare and Education","score":0.00139999995008111,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/whole-slide-images","display_name":"whole slide images","score":0.8349999785423279},{"id":"https://openalex.org/keywords/use-of-evidence","display_name":"evidence use","score":0.878000020980835},{"id":"https://openalex.org/keywords/prior-knowledge","display_name":"prior knowledge","score":0.6819999814033508},{"id":"https://openalex.org/keywords/lymph-node-metastasis-prediction","display_name":"lymph node metastasis prediction","score":0.7990000247955322}],"concepts":[{"id":"https://openalex.org/C108650721","wikidata":"https://www.wikidata.org/wiki/Q1783253","display_name":"Counterfactual thinking","level":2,"score":0.8331000208854675},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.8130999803543091},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6389999985694885},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.6187000274658203},{"id":"https://openalex.org/C199521495","wikidata":"https://www.wikidata.org/wiki/Q181487","display_name":"Audit","level":2,"score":0.5568000078201294},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5250999927520752},{"id":"https://openalex.org/C115961682","wikidata":"https://www.wikidata.org/wiki/Q860623","display_name":"Image (mathematics)","level":2,"score":0.5131000280380249},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.42879998683929443},{"id":"https://openalex.org/C2781249084","wikidata":"https://www.wikidata.org/wiki/Q908656","display_name":"Preference","level":2,"score":0.4115999937057495},{"id":"https://openalex.org/C2780009758","wikidata":"https://www.wikidata.org/wiki/Q6804172","display_name":"Measure (data warehouse)","level":2,"score":0.38449999690055847},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.3440000116825104},{"id":"https://openalex.org/C177148314","wikidata":"https://www.wikidata.org/wiki/Q170084","display_name":"Generalization","level":2,"score":0.33149999380111694},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.325300008058548},{"id":"https://openalex.org/C2776502983","wikidata":"https://www.wikidata.org/wiki/Q690182","display_name":"Contrast (vision)","level":2,"score":0.3091999888420105},{"id":"https://openalex.org/C130440534","wikidata":"https://www.wikidata.org/wiki/Q14946528","display_name":"Conflation","level":2,"score":0.2955000102519989},{"id":"https://openalex.org/C77350462","wikidata":"https://www.wikidata.org/wiki/Q1125472","display_name":"Confounding","level":2,"score":0.28760001063346863},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.2685000002384186}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2609.32185","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.32185","pdf_url":"https://arxiv.org/pdf/2609.32185","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.2609.32185","doi":"https://doi.org/10.48550/arxiv.2609.32185","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2609.32185","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2609.32185","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.32185","pdf_url":"https://arxiv.org/pdf/2609.32185","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":[{"score":0.4043000042438507,"id":"https://metadata.un.org/sdg/3","display_name":"Good health and well-being"}],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7214780475.pdf","grobid_xml":"https://content.openalex.org/works/W7214780475.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W7170071148","https://openalex.org/W7223012147","https://openalex.org/W7125704730","https://openalex.org/W7139038381","https://openalex.org/W7168352138","https://openalex.org/W7221074449","https://openalex.org/W7212385991","https://openalex.org/W7167511836","https://openalex.org/W4409710058","https://openalex.org/W7203743890"],"abstract_inverted_index":{"Pathology":[0],"vision-language":[1],"models":[2,50,234],"(VLMs)":[3],"are":[4],"conventionally":[5],"evaluated":[6,233],"by":[7,69,235],"accuracy,":[8],"but":[9],"accuracy":[10],"alone":[11],"does":[12],"not":[13],"measure":[14],"evidence":[15,157,203],"use:":[16],"it":[17],"may":[18],"conflate":[19],"dataset":[20],"contamination,":[21],"prior":[22],"knowledge,":[23],"and":[24,94,108,122,140,154,191,219,228,237,245],"image":[25,59,156,202],"evidence.":[26],"In":[27],"a":[28,86,129],"motivating":[29],"study":[30],"of":[31,161,167,226],"lymph-node":[32],"metastasis":[33],"prediction,":[34],"we":[35,83,126],"found":[36],"that":[37,196],"most":[38],"public":[39],"pathology":[40],"VLMs":[41],"showed":[42],"minimal":[43],"differences":[44],"when":[45],"changing":[46],"from":[47,136,201],"feeding":[48],"the":[49,65,137,151,159,205,211,216,243],"with":[51,106,171,208],"whole-slide":[52],"images,":[53],"an":[54],"annotated":[55],"lesion,":[56],"or":[57,184],"no":[58],"at":[60,78],"all.":[61],"To":[62],"better":[63],"understand":[64],"specific":[66],"features":[67,186],"leveraged":[68],"these":[70,80],"models,":[71],"this":[72],"paper":[73],"presents":[74],"two":[75,168],"contributions":[76],"aimed":[77],"disentangling":[79],"factors.":[81],"First,":[82],"present":[84],"CleanSlide,":[85,206],"TCGA-based":[87],"VQA":[88],"benchmark":[89,244],"designed":[90],"to":[91],"eliminate":[92],"image-":[93],"question-side":[95],"contamination.":[96],"It":[97],"contains":[98],"149K":[99],"audited":[100,114],"multiple-choice":[101],"questions":[102],"over":[103,132],"9,985":[104],"slides,":[105],"patient-":[107],"tissue-source-disjoint":[109],"splits.":[110],"Every":[111],"question":[112,139],"is":[113],"for":[115,144,182,210],"option":[116],"shortcuts,":[117],"stem":[118],"leakage,":[119],"cross-split":[120],"duplication,":[121],"blind":[123],"solvability.":[124],"Second,":[125],"propose":[127],"Pair-DPO,":[128],"preference":[130],"loss":[131],"counterfactual":[133],"slide":[134],"pairs":[135],"same":[138],"source.":[141],"By":[142],"controlling":[143],"shared":[145],"confounding":[146],"factors,":[147],"Pair-DPO":[148],"cancels":[149],"out":[150],"question-attributable":[152],"signal":[153],"leaves":[155],"as":[158,188],"source":[160],"preference.":[162],"Specifically,":[163],"each":[164],"pair":[165],"consists":[166],"real":[169],"slides":[170],"opposite,":[172],"verified":[173],"findings,":[174],"introducing":[175],"neither":[176],"editing":[177],"artifacts":[178],"nor":[179],"unverified":[180],"labels":[181],"diffuse":[183],"graded":[185],"such":[187],"invasion,":[189],"necrosis,":[190],"tumor":[192],"grade.":[193],"Experiments":[194],"show":[195],"our":[197,222],"method":[198,223],"gains":[199],"15.29%":[200],"on":[204],"compared":[207],"2.81%":[209],"best":[212],"published":[213],"model.":[214],"On":[215],"external":[217],"CPTAC":[218],"BCNB":[220],"cohorts,":[221],"achieves":[224],"accuracies":[225],"57.6%":[227],"59.0%,":[229],"outperforming":[230],"all":[231],"other":[232],"9.7%":[236],"3.4%,":[238],"respectively.":[239],"We":[240],"will":[241],"release":[242],"code.":[246]},"counts_by_year":[],"updated_date":"2026-10-28T10:33:24.094632","created_date":"2026-09-30T00:00:00"}