{"id":"https://openalex.org/W7167081638","doi":"https://doi.org/10.48550/arxiv.2607.00125","title":"Decompose, Compare, and Decide: Multimodal LLMs are Implicit Few-Shot Learners","display_name":"Decompose, Compare, and Decide: Multimodal LLMs are Implicit Few-Shot Learners","publication_year":2026,"publication_date":"2026-06-30","ids":{"openalex":"https://openalex.org/W7167081638","doi":"https://doi.org/10.48550/arxiv.2607.00125"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2607.00125","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2607.00125","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,"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.2607.00125","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5139861282","display_name":"Yunhan Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Yunhan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5093382101","display_name":"Eshika Khandelwal","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Khandelwal, Eshika","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134157762","display_name":"Edson Araujo","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Araujo, Edson","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5030078138","display_name":"Walid Bousselham","orcid":"https://orcid.org/0000-0002-9684-523X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Bousselham, Walid","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139897318","display_name":"Nina Shvetsova","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shvetsova, Nina","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5139883453","display_name":"Hilde Kuehne","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Kuehne, Hilde","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":0.574999988079071,"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":0.574999988079071,"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":0.09459999948740005,"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/T10028","display_name":"Topic Modeling","score":0.08649999648332596,"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/pairwise-comparison","display_name":"Pairwise comparison","score":0.5837000012397766},{"id":"https://openalex.org/keywords/set","display_name":"Set (abstract data type)","score":0.5764999985694885},{"id":"https://openalex.org/keywords/image","display_name":"Image (mathematics)","score":0.5587000250816345},{"id":"https://openalex.org/keywords/task","display_name":"Task (project management)","score":0.490200012922287},{"id":"https://openalex.org/keywords/similarity","display_name":"Similarity (geometry)","score":0.4794999957084656},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.4724999964237213},{"id":"https://openalex.org/keywords/code","display_name":"Code (set theory)","score":0.4172999858856201},{"id":"https://openalex.org/keywords/pattern-recognition","display_name":"Pattern recognition (psychology)","score":0.4043999910354614}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6796000003814697},{"id":"https://openalex.org/C184898388","wikidata":"https://www.wikidata.org/wiki/Q1435712","display_name":"Pairwise comparison","level":2,"score":0.5837000012397766},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.5764999985694885},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5734999775886536},{"id":"https://openalex.org/C115961682","wikidata":"https://www.wikidata.org/wiki/Q860623","display_name":"Image (mathematics)","level":2,"score":0.5587000250816345},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5097000002861023},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.490200012922287},{"id":"https://openalex.org/C103278499","wikidata":"https://www.wikidata.org/wiki/Q254465","display_name":"Similarity (geometry)","level":3,"score":0.4794999957084656},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.4724999964237213},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.4172999858856201},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.4043999910354614},{"id":"https://openalex.org/C2780586882","wikidata":"https://www.wikidata.org/wiki/Q7520643","display_name":"Simple (philosophy)","level":2,"score":0.40400001406669617},{"id":"https://openalex.org/C79581498","wikidata":"https://www.wikidata.org/wiki/Q1367530","display_name":"Suite","level":2,"score":0.3984000086784363},{"id":"https://openalex.org/C66905080","wikidata":"https://www.wikidata.org/wiki/Q17005494","display_name":"Binary classification","level":3,"score":0.36660000681877136},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.35280001163482666},{"id":"https://openalex.org/C48372109","wikidata":"https://www.wikidata.org/wiki/Q3913","display_name":"Binary number","level":2,"score":0.34220001101493835},{"id":"https://openalex.org/C75294576","wikidata":"https://www.wikidata.org/wiki/Q5165192","display_name":"Contextual image classification","level":3,"score":0.33309999108314514},{"id":"https://openalex.org/C175154964","wikidata":"https://www.wikidata.org/wiki/Q380077","display_name":"Task analysis","level":3,"score":0.3091999888420105},{"id":"https://openalex.org/C64543145","wikidata":"https://www.wikidata.org/wiki/Q162942","display_name":"Intersection (aeronautics)","level":2,"score":0.28870001435279846},{"id":"https://openalex.org/C58489278","wikidata":"https://www.wikidata.org/wiki/Q1172284","display_name":"Data set","level":2,"score":0.28769999742507935},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.2867000102996826},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.28380000591278076}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2607.00125","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2607.00125","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,"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.2607.00125","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2607.00125","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,"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":0.7081191539764404,"id":"https://metadata.un.org/sdg/4","display_name":"Quality Education"}],"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],"Models":[3],"(MLLMs)":[4],"have":[5],"demonstrated":[6],"remarkable":[7],"abilities":[8],"when":[9],"analyzing":[10],"images,":[11],"yet":[12,31],"translating":[13],"these":[14],"capabilities":[15],"to":[16,38,89,102,113,118,141],"few-shot":[17,42,55,169,190,196],"image":[18,62,76,80,117],"classification":[19,56],"remains":[20],"challenging.":[21],"To":[22],"bridge":[23],"this":[24,124],"gap,":[25],"we":[26,129],"present":[27],"DeCoDe,":[28],"a":[29,58,68,74,78,82,110,158],"simple":[30,181],"effective":[32],"technique":[33,183],"that":[34,131,178],"enables":[35],"off-the-shelf":[36,186],"MLLMs":[37,187],"act":[39],"as":[40,57,109,137],"strong":[41],"classifiers":[43],"without":[44],"any":[45],"additional":[46,133],"training.":[47],"Our":[48,147],"approach":[49],"builds":[50],"on":[51,157,198],"the":[52,65,85,92,96,115,119,138,142,179],"idea":[53],"of":[54,60,70,153,160],"set":[59,69],"pairwise":[61],"comparisons,":[63],"decomposing":[64],"task":[66],"into":[67,188],"binary":[71],"decisions.":[72],"Given":[73],"query":[75,116],"and":[77,165,201],"support":[79],"from":[81],"candidate":[83],"class,":[84],"MLLM":[86],"is":[87,106,205],"prompted":[88],"decide":[90],"whether":[91],"two":[93],"images":[94],"depict":[95],"same":[97],"class.":[98,122],"The":[99,175],"logit":[100],"corresponding":[101],"an":[103,150],"affirmative":[104],"response":[105],"then":[107],"used":[108],"similarity":[111],"score":[112],"assign":[114],"most":[120],"likely":[121],"While":[123],"already":[125],"yields":[126],"good":[127],"results,":[128],"show":[130,177],"providing":[132],"high-level":[134],"information,":[135],"such":[136],"data":[139],"domain,":[140],"model":[143],"further":[144],"improves":[145],"performance.":[146],"evaluation":[148],"provides":[149],"extensive":[151],"analysis":[152],"various":[154],"inference":[155],"variants":[156],"suite":[159],"twelve":[161],"datasets,":[162],"six":[163,166],"established":[164],"newly":[167],"curated":[168],"benchmarks":[170],"spanning":[171],"across":[172],"diverse":[173],"domains.":[174,203],"results":[176],"proposed":[180],"decomposition":[182],"can":[184],"turn":[185],"powerful":[189],"learners,":[191],"significantly":[192],"outperforming":[193],"current":[194],"state-of-the-art":[195],"methods":[197],"both":[199],"standard":[200],"novel":[202],"Code":[204],"available":[206],"at":[207],"https://github.com/yunhanwang1105/DeCoDe.":[208]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2026-07-03T00:00:00"}