{"id":"https://openalex.org/W7131620433","doi":"https://doi.org/10.48550/arxiv.2602.21372","title":"The Mean is the Mirage: Entropy-Adaptive Model Merging under Heterogeneous Domain Shifts in Medical Imaging","display_name":"The Mean is the Mirage: Entropy-Adaptive Model Merging under Heterogeneous Domain Shifts in Medical Imaging","publication_year":2026,"publication_date":"2026-02-24","ids":{"openalex":"https://openalex.org/W7131620433","doi":"https://doi.org/10.48550/arxiv.2602.21372"},"language":null,"primary_location":{"id":"pmh:doi:10.48550/arxiv.2602.21372","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"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":"Article"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":null,"any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5068852913","display_name":"Sameer Ambekar","orcid":"https://orcid.org/0000-0002-8650-3180"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ambekar, Sameer","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5048531338","display_name":"Reza Nasirigerdeh","orcid":"https://orcid.org/0000-0002-2283-5792"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Nasirigerdeh, Reza","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5029617163","display_name":"Peter J. Sch\u00fcffler","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Schuffler, Peter J.","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5042531662","display_name":"Lina Felsner","orcid":"https://orcid.org/0000-0001-7695-2612"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Felsner, Lina","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5070331938","display_name":"Daniel M. Lang","orcid":"https://orcid.org/0000-0003-0274-9069"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lang, Daniel M.","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5126876339","display_name":"Julia A. Schnabel","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Schnabel, Julia A.","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":6,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.8741000294685364,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.8741000294685364,"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/T13702","display_name":"Machine Learning in Healthcare","score":0.01899999938905239,"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/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.016599999740719795,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/encoder","display_name":"Encoder","score":0.6154999732971191},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.6079000234603882},{"id":"https://openalex.org/keywords/generalization","display_name":"Generalization","score":0.5400999784469604},{"id":"https://openalex.org/keywords/medical-imaging","display_name":"Medical imaging","score":0.4742000102996826},{"id":"https://openalex.org/keywords/classifier","display_name":"Classifier (UML)","score":0.453000009059906},{"id":"https://openalex.org/keywords/domain","display_name":"Domain (mathematical analysis)","score":0.45019999146461487}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7305999994277954},{"id":"https://openalex.org/C118505674","wikidata":"https://www.wikidata.org/wiki/Q42586063","display_name":"Encoder","level":2,"score":0.6154999732971191},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.6079000234603882},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5835999846458435},{"id":"https://openalex.org/C177148314","wikidata":"https://www.wikidata.org/wiki/Q170084","display_name":"Generalization","level":2,"score":0.5400999784469604},{"id":"https://openalex.org/C31601959","wikidata":"https://www.wikidata.org/wiki/Q931309","display_name":"Medical imaging","level":2,"score":0.4742000102996826},{"id":"https://openalex.org/C95623464","wikidata":"https://www.wikidata.org/wiki/Q1096149","display_name":"Classifier (UML)","level":2,"score":0.453000009059906},{"id":"https://openalex.org/C36503486","wikidata":"https://www.wikidata.org/wiki/Q11235244","display_name":"Domain (mathematical analysis)","level":2,"score":0.45019999146461487},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4163999855518341},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.3840000033378601},{"id":"https://openalex.org/C205606062","wikidata":"https://www.wikidata.org/wiki/Q5249645","display_name":"Decoupling (probability)","level":2,"score":0.3684000074863434},{"id":"https://openalex.org/C101738243","wikidata":"https://www.wikidata.org/wiki/Q786435","display_name":"Autoencoder","level":3,"score":0.3580000102519989},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.33640000224113464},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3312999904155731},{"id":"https://openalex.org/C2777472644","wikidata":"https://www.wikidata.org/wiki/Q16968992","display_name":"Approximate inference","level":3,"score":0.2849000096321106},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.26339998841285706}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:doi:10.48550/arxiv.2602.21372","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"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":"Article"},{"id":"doi:10.48550/arxiv.2602.21372","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2602.21372","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":"article"}],"best_oa_location":{"id":"pmh:doi:10.48550/arxiv.2602.21372","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"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":"Article"},"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":{"Model":[0],"merging":[1,96,120,123],"under":[2,103],"unseen":[3,48],"test-time":[4],"distribution":[5],"shifts":[6],"often":[7],"renders":[8],"naive":[9],"strategies,":[10],"such":[11],"as":[12],"mean":[13,95],"averaging":[14],"unreliable.":[15],"This":[16],"challenge":[17],"is":[18,97],"especially":[19],"acute":[20],"in":[21,54],"medical":[22,138],"imaging,":[23],"where":[24],"models":[25,36],"are":[26,157],"fine-tuned":[27],"locally":[28],"at":[29,46,163],"clinics":[30],"on":[31],"private":[32],"data,":[33],"producing":[34],"domain-specific":[35],"that":[37,77],"differ":[38],"by":[39,113],"scanner,":[40],"protocol,":[41],"and":[42,58,101,117,139,151],"population.":[43],"When":[44],"deployed":[45],"an":[47,71],"clinical":[49],"site,":[50],"test":[51],"cases":[52],"arrive":[53],"unlabeled,":[55],"non-i.i.d.":[56],"batches,":[57],"the":[59,115,167],"model":[60,82],"must":[61],"adapt":[62],"immediately":[63],"without":[64],"labels.":[65],"In":[66],"this":[67],"work,":[68],"we":[69,108],"introduce":[70],"entropy-adaptive,":[72],"fully":[73],"online":[74],"model-merging":[75],"method":[76,129],"yields":[78],"a":[79],"batch-specific":[80],"merged":[81],"via":[83],"only":[84],"forward":[85],"passes,":[86],"effectively":[87],"leveraging":[88],"target":[89],"information.":[90],"We":[91,125],"further":[92],"demonstrate":[93],"why":[94],"prone":[98],"to":[99],"failure":[100],"misaligned":[102],"heterogeneous":[104],"domain":[105],"shifts.":[106],"Next,":[107],"mitigate":[109],"encoder":[110,116],"classifier":[111],"mismatch":[112],"decoupling":[114],"classification":[118,143],"head,":[119],"with":[121,130],"separate":[122],"coefficients.":[124],"extensively":[126],"evaluate":[127],"our":[128,170],"state-of-the-art":[131],"baselines":[132],"using":[133],"two":[134],"backbones":[135],"across":[136,148],"nine":[137],"natural-domain":[140],"generalization":[141],"image":[142],"datasets,":[144],"showing":[145],"consistent":[146],"gains":[147,156],"standard":[149],"evaluation":[150],"challenging":[152],"scenarios.":[153],"These":[154],"performance":[155],"achieved":[158],"while":[159],"retaining":[160],"single-model":[161],"inference":[162],"test-time,":[164],"thereby":[165],"demonstrating":[166],"effectiveness":[168],"of":[169],"method.":[171]},"counts_by_year":[],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2026-02-27T00:00:00"}