{"id":"https://openalex.org/W7138299503","doi":"https://doi.org/10.48550/arxiv.2603.13341","title":"Mind the Discriminability Trap in Source-Free Cross-domain Few-shot Learning","display_name":"Mind the Discriminability Trap in Source-Free Cross-domain Few-shot Learning","publication_year":2026,"publication_date":"2026-03-07","ids":{"openalex":"https://openalex.org/W7138299503","doi":"https://doi.org/10.48550/arxiv.2603.13341"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2603.13341","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.13341","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":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.2603.13341","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5129652573","display_name":"Zhenyu Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Zhenyu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129681083","display_name":"Yixiong Zou","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zou, Yixiong","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129710056","display_name":"Yuhua Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Yuhua","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129683157","display_name":"Ruixuan Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Ruixuan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5129691005","display_name":"Guangyao Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Guangyao","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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.8809000253677368,"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":0.8809000253677368,"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":0.09669999778270721,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.007899999618530273,"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/focus","display_name":"Focus (optics)","score":0.6614999771118164},{"id":"https://openalex.org/keywords/set","display_name":"Set (abstract data type)","score":0.6022999882698059},{"id":"https://openalex.org/keywords/interpretation","display_name":"Interpretation (philosophy)","score":0.5523999929428101},{"id":"https://openalex.org/keywords/visual-learning","display_name":"Visual learning","score":0.5332000255584717},{"id":"https://openalex.org/keywords/code","display_name":"Code (set theory)","score":0.4986000061035156},{"id":"https://openalex.org/keywords/modalities","display_name":"Modalities","score":0.4941999912261963},{"id":"https://openalex.org/keywords/modality","display_name":"Modality (human\u2013computer interaction)","score":0.43209999799728394},{"id":"https://openalex.org/keywords/phenomenon","display_name":"Phenomenon","score":0.39500001072883606}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6618000268936157},{"id":"https://openalex.org/C192209626","wikidata":"https://www.wikidata.org/wiki/Q190909","display_name":"Focus (optics)","level":2,"score":0.6614999771118164},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.6022999882698059},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5857999920845032},{"id":"https://openalex.org/C527412718","wikidata":"https://www.wikidata.org/wiki/Q855395","display_name":"Interpretation (philosophy)","level":2,"score":0.5523999929428101},{"id":"https://openalex.org/C2779321571","wikidata":"https://www.wikidata.org/wiki/Q7936605","display_name":"Visual learning","level":2,"score":0.5332000255584717},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.4986000061035156},{"id":"https://openalex.org/C2779903281","wikidata":"https://www.wikidata.org/wiki/Q6888026","display_name":"Modalities","level":2,"score":0.4941999912261963},{"id":"https://openalex.org/C2780226545","wikidata":"https://www.wikidata.org/wiki/Q6888030","display_name":"Modality (human\u2013computer interaction)","level":2,"score":0.43209999799728394},{"id":"https://openalex.org/C50335755","wikidata":"https://www.wikidata.org/wiki/Q483247","display_name":"Phenomenon","level":2,"score":0.39500001072883606},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.3483000099658966},{"id":"https://openalex.org/C2986089797","wikidata":"https://www.wikidata.org/wiki/Q6501338","display_name":"Visual attention","level":3,"score":0.3276999890804291},{"id":"https://openalex.org/C121099081","wikidata":"https://www.wikidata.org/wiki/Q665580","display_name":"Trap (plumbing)","level":2,"score":0.3125999867916107},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3075000047683716},{"id":"https://openalex.org/C2776502983","wikidata":"https://www.wikidata.org/wiki/Q690182","display_name":"Contrast (vision)","level":2,"score":0.28630000352859497},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.28540000319480896},{"id":"https://openalex.org/C178253425","wikidata":"https://www.wikidata.org/wiki/Q162668","display_name":"Visual perception","level":3,"score":0.2840999960899353},{"id":"https://openalex.org/C66024118","wikidata":"https://www.wikidata.org/wiki/Q1122506","display_name":"Computational model","level":2,"score":0.28360000252723694},{"id":"https://openalex.org/C2778251979","wikidata":"https://www.wikidata.org/wiki/Q7936617","display_name":"Visual processing","level":3,"score":0.2524000108242035}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2603.13341","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.13341","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":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.2603.13341","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.13341","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"sustainable_development_goals":[{"score":0.7490342259407043,"id":"https://metadata.un.org/sdg/10","display_name":"Reduced inequalities"}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Source-Free":[0],"Cross-Domain":[1],"Few-Shot":[2],"Learning":[3],"(SF-CDFSL)":[4],"focuses":[5],"on":[6,154,178,202],"fine-tuning":[7,87],"with":[8,88],"limited":[9],"training":[10],"data":[11],"from":[12],"target":[13],"domains":[14],"(e.g.,":[15],"medical":[16],"or":[17],"satellite":[18],"images),":[19],"where":[20,105],"Vision-Language":[21],"Models":[22],"(VLMs)":[23],"such":[24],"as":[25,129],"CLIP":[26],"and":[27,74,80,100,149,194,209,214],"SigLIP":[28],"have":[29],"shown":[30],"promising":[31],"results.":[32,225],"Current":[33],"works":[34],"in":[35,47,118],"traditional":[36],"visual":[37,42,97,125,170,193],"models":[38],"suggest":[39],"that":[40,53,86,123,132,219],"improving":[41],"discriminability":[43,56],"enhances":[44],"performance.":[45,152],"However,":[46,120],"VLM-based":[48],"SF-CDFSL":[49],"tasks,":[50],"we":[51,64,121,157,167,183,220],"find":[52,122],"\\textbf{strengthening":[54],"visual-modal":[55],"actually":[57],"suppresses":[58],"VLMs'":[59],"performance}.":[60],"In":[61],"this":[62,69,155,164],"paper,":[63],"aim":[65],"to":[66,136,162,172,176,189],"delve":[67],"into":[68],"phenomenon":[70],"for":[71,111],"an":[72,160],"interpretation":[73],"a":[75,96,101,130],"solution.":[76],"By":[77],"both":[78],"theoretical":[79],"experimental":[81],"proofs,":[82],"our":[83],"study":[84],"reveals":[85],"the":[89,106,113,124,134,141,146,151,169,174,179,185,192,198],"typical":[90],"cross-entropy":[91],"loss":[92],"($\\mathcal{L}_{\\mathrm{vlm}}$)":[93],"inherently":[94],"includes":[95],"learning":[98,103,126,171],"part":[99,108],"cross-modal":[102,107,142,147,180],"part,":[104,143],"is":[109,227],"crucial":[110],"rectifying":[112],"heavily":[114],"disrupted":[115],"modality":[116],"misalignment":[117],"SF-CDFSL.":[119],"essentially":[127],"acts":[128],"shortcut":[131],"encourages":[133],"model":[135,175],"reduce":[137],"$\\mathcal{L}_{\\mathrm{vlm}}$":[138],"without":[139],"considering":[140],"therefore":[144],"hindering":[145],"alignment":[148],"harming":[150],"Based":[153],"interpretation,":[156],"further":[158],"propose":[159],"approach":[161],"address":[163],"problem:":[165],"first,":[166],"perturb":[168],"guide":[173],"focus":[177],"alignment.":[181],"Then,":[182],"use":[184],"visual-text":[186],"semantic":[187],"relationships":[188],"gradually":[190],"align":[191],"textual":[195],"modalities":[196],"during":[197],"fine-tuning.":[199],"Extensive":[200],"experiments":[201],"various":[203],"settings,":[204],"backbones":[205],"(CLIP,":[206],"SigLip,":[207],"PE-Core),":[208],"tasks":[210],"(4":[211],"CDFSL":[212],"datasets":[213],"11":[215],"FSL":[216],"datasets)":[217],"show":[218],"consistently":[221],"set":[222],"new":[223],"state-of-the-art":[224],"Code":[226],"available":[228],"at":[229],"https://github.com/zhenyuZ-HUST/CVPR26-Mind-the-Discriminability-Trap.":[230]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-03-18T00:00:00"}