{"id":"https://openalex.org/W7124195832","doi":"https://doi.org/10.48550/arxiv.2601.08151","title":"Where Does Vision Meet Language? Understanding and Refining Visual Fusion in MLLMs via Contrastive Attention","display_name":"Where Does Vision Meet Language? Understanding and Refining Visual Fusion in MLLMs via Contrastive Attention","publication_year":2026,"publication_date":"2026-01-13","ids":{"openalex":"https://openalex.org/W7124195832","doi":"https://doi.org/10.48550/arxiv.2601.08151"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2601.08151","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2601.08151","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":"article"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2601.08151","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":null,"display_name":"Song, Shezheng","orcid":null},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Song, Shezheng","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5123053332","display_name":"Shasha Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Shasha","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5123043458","display_name":"Jie Yu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yu, Jie","raw_affiliation_strings":[],"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":3,"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":0.8813999891281128,"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":0.8813999891281128,"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/T11148","display_name":"Language, Metaphor, and Cognition","score":0.010499999858438969,"subfield":{"id":"https://openalex.org/subfields/3205","display_name":"Experimental and Cognitive Psychology"},"field":{"id":"https://openalex.org/fields/32","display_name":"Psychology"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}},{"id":"https://openalex.org/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.009700000286102295,"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/masking","display_name":"Masking (illustration)","score":0.5371000170707703},{"id":"https://openalex.org/keywords/fusion","display_name":"Fusion","score":0.41920000314712524},{"id":"https://openalex.org/keywords/transformation","display_name":"Transformation (genetics)","score":0.4068000018596649},{"id":"https://openalex.org/keywords/visual-masking","display_name":"Visual masking","score":0.37389999628067017},{"id":"https://openalex.org/keywords/bridge","display_name":"Bridge (graph theory)","score":0.37070000171661377},{"id":"https://openalex.org/keywords/sensor-fusion","display_name":"Sensor fusion","score":0.366100013256073},{"id":"https://openalex.org/keywords/visual-reasoning","display_name":"Visual reasoning","score":0.34389999508857727},{"id":"https://openalex.org/keywords/noise","display_name":"Noise (video)","score":0.33640000224113464}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7153000235557556},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5390999913215637},{"id":"https://openalex.org/C2777402240","wikidata":"https://www.wikidata.org/wiki/Q6783436","display_name":"Masking (illustration)","level":2,"score":0.5371000170707703},{"id":"https://openalex.org/C158525013","wikidata":"https://www.wikidata.org/wiki/Q2593739","display_name":"Fusion","level":2,"score":0.41920000314712524},{"id":"https://openalex.org/C204241405","wikidata":"https://www.wikidata.org/wiki/Q461499","display_name":"Transformation (genetics)","level":3,"score":0.4068000018596649},{"id":"https://openalex.org/C2779200073","wikidata":"https://www.wikidata.org/wiki/Q18395575","display_name":"Visual masking","level":4,"score":0.37389999628067017},{"id":"https://openalex.org/C100776233","wikidata":"https://www.wikidata.org/wiki/Q2532492","display_name":"Bridge (graph theory)","level":2,"score":0.37070000171661377},{"id":"https://openalex.org/C33954974","wikidata":"https://www.wikidata.org/wiki/Q486494","display_name":"Sensor fusion","level":2,"score":0.366100013256073},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.3605000078678131},{"id":"https://openalex.org/C2777508537","wikidata":"https://www.wikidata.org/wiki/Q7936620","display_name":"Visual reasoning","level":2,"score":0.34389999508857727},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.33640000224113464},{"id":"https://openalex.org/C112313634","wikidata":"https://www.wikidata.org/wiki/Q7886648","display_name":"Complement (music)","level":5,"score":0.32359999418258667},{"id":"https://openalex.org/C107457646","wikidata":"https://www.wikidata.org/wiki/Q207434","display_name":"Human\u2013computer interaction","level":1,"score":0.3160000145435333},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.3138999938964844},{"id":"https://openalex.org/C66024118","wikidata":"https://www.wikidata.org/wiki/Q1122506","display_name":"Computational model","level":2,"score":0.29420000314712524},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.2815000116825104},{"id":"https://openalex.org/C178253425","wikidata":"https://www.wikidata.org/wiki/Q162668","display_name":"Visual perception","level":3,"score":0.2786000072956085},{"id":"https://openalex.org/C2776817677","wikidata":"https://www.wikidata.org/wiki/Q4839818","display_name":"Backward masking","level":3,"score":0.27070000767707825},{"id":"https://openalex.org/C60044698","wikidata":"https://www.wikidata.org/wiki/Q1283324","display_name":"Refining (metallurgy)","level":2,"score":0.2621000111103058},{"id":"https://openalex.org/C157657479","wikidata":"https://www.wikidata.org/wiki/Q2367247","display_name":"Closed captioning","level":3,"score":0.2612000107765198},{"id":"https://openalex.org/C69744172","wikidata":"https://www.wikidata.org/wiki/Q860822","display_name":"Image fusion","level":3,"score":0.26030001044273376},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.2540999948978424},{"id":"https://openalex.org/C26760741","wikidata":"https://www.wikidata.org/wiki/Q160402","display_name":"Perception","level":2,"score":0.25099998712539673},{"id":"https://openalex.org/C36464697","wikidata":"https://www.wikidata.org/wiki/Q451553","display_name":"Visualization","level":2,"score":0.2500999867916107}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2601.08151","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2601.08151","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":"doi:10.48550/arxiv.2601.08151","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2601.08151","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":"article"},"sustainable_development_goals":[{"id":"https://metadata.un.org/sdg/4","score":0.80437833070755,"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],"achieved":[6],"remarkable":[7],"progress":[8],"in":[9],"vision-language":[10],"understanding,":[11],"yet":[12],"how":[13,39],"they":[14],"internally":[15],"integrate":[16],"visual":[17,72],"and":[18,63,86,120,133,138],"textual":[19],"information":[20],"remains":[21],"poorly":[22],"understood.":[23],"To":[24],"bridge":[25],"this":[26],"gap,":[27],"we":[28,80,106],"perform":[29],"a":[30,67,108],"systematic":[31],"layer-wise":[32,83],"masking":[33],"analysis":[34,137],"across":[35,60,130],"multiple":[36],"architectures,":[37],"revealing":[38],"visual-text":[40],"fusion":[41,49,119],"evolves":[42],"within":[43],"MLLMs.":[44],"The":[45],"results":[46],"show":[47],"that":[48,113,140],"emerges":[50],"at":[51],"several":[52],"specific":[53],"layers":[54,122],"rather":[55],"than":[56],"being":[57],"uniformly":[58],"distributed":[59],"the":[61,115,141],"network,":[62],"certain":[64],"models":[65,114],"exhibit":[66],"late-stage":[68],"\"review\"":[69],"phenomenon":[70],"where":[71],"signals":[73],"are":[74],"reactivated":[75],"before":[76],"output":[77],"generation.":[78],"Besides,":[79],"further":[81],"analyze":[82],"attention":[84,98,111,126],"evolution":[85],"observe":[87],"persistent":[88],"high-attention":[89],"noise":[90],"on":[91,99],"irrelevant":[92],"regions,":[93],"along":[94],"with":[95],"gradually":[96],"increasing":[97],"text-aligned":[100],"areas.":[101],"Guided":[102],"by":[103],"these":[104],"insights,":[105],"introduce":[107],"training-free":[109],"contrastive":[110],"framework":[112],"transformation":[116],"between":[117],"early":[118],"final":[121],"to":[123],"highlight":[124],"meaningful":[125],"shifts.":[127],"Extensive":[128],"experiments":[129],"various":[131],"MLLMs":[132],"benchmarks":[134],"validate":[135],"our":[136],"demonstrate":[139],"proposed":[142],"approach":[143],"improves":[144],"multimodal":[145],"reasoning":[146],"performance.":[147],"Code":[148],"will":[149],"be":[150],"released.":[151]},"counts_by_year":[],"updated_date":"2026-03-25T23:56:10.502304","created_date":"2026-01-15T00:00:00"}