{"id":"https://openalex.org/W4417093151","doi":"https://doi.org/10.48550/arxiv.2505.10917","title":"VISTA: Enhancing Vision-Text Alignment in MLLMs via Cross-Modal Mutual Information Maximization","display_name":"VISTA: Enhancing Vision-Text Alignment in MLLMs via Cross-Modal Mutual Information Maximization","publication_year":2025,"publication_date":"2025-05-16","ids":{"openalex":"https://openalex.org/W4417093151","doi":"https://doi.org/10.48550/arxiv.2505.10917"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2505.10917","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2505.10917","pdf_url":"https://arxiv.org/pdf/2505.10917","source":{"id":"https://openalex.org/S4393918464","display_name":"ArXiv.org","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"type":"preprint","indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2505.10917","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5101583267","display_name":"Mingxiao Li","orcid":"https://orcid.org/0000-0003-0779-4432"},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Li, Mingxiao","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5103243484","display_name":"Su Na","orcid":"https://orcid.org/0000-0003-2215-5575"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Su, Na","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102199800","display_name":"Fang Qu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Qu, Fang","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Zhong, Zhizhou","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhong, Zhizhou","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100324487","display_name":"Ziyang Chen","orcid":"https://orcid.org/0000-0002-8564-9735"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Ziyang","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101903323","display_name":"Yuan Li","orcid":"https://orcid.org/0000-0003-1500-8884"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Yuan","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5110579339","display_name":"Zhaopeng Tu","orcid":"https://orcid.org/0000-0002-7900-6055"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tu, Zhaopeng","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5100371535","display_name":"Xiaolong Li","orcid":"https://orcid.org/0000-0001-7493-2650"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Xiaolong","raw_affiliation_strings":[],"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":8,"corresponding_author_ids":["https://openalex.org/A5101583267"],"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":0.9071000218391418,"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":0.9071000218391418,"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":0.03739999979734421,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.010099999606609344,"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/mutual-information","display_name":"Mutual information","score":0.6597999930381775},{"id":"https://openalex.org/keywords/maximization","display_name":"Maximization","score":0.6115000247955322},{"id":"https://openalex.org/keywords/benchmark","display_name":"Benchmark (surveying)","score":0.5249000191688538},{"id":"https://openalex.org/keywords/face","display_name":"Face (sociological concept)","score":0.4778999984264374},{"id":"https://openalex.org/keywords/information-loss","display_name":"Information loss","score":0.47440001368522644},{"id":"https://openalex.org/keywords/modality","display_name":"Modality (human\u2013computer interaction)","score":0.42590001225471497},{"id":"https://openalex.org/keywords/baseline","display_name":"Baseline (sea)","score":0.4212999939918518},{"id":"https://openalex.org/keywords/modalities","display_name":"Modalities","score":0.4002000093460083}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7872999906539917},{"id":"https://openalex.org/C152139883","wikidata":"https://www.wikidata.org/wiki/Q252973","display_name":"Mutual information","level":2,"score":0.6597999930381775},{"id":"https://openalex.org/C2776330181","wikidata":"https://www.wikidata.org/wiki/Q18358244","display_name":"Maximization","level":2,"score":0.6115000247955322},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.5249000191688538},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5164999961853027},{"id":"https://openalex.org/C2779304628","wikidata":"https://www.wikidata.org/wiki/Q3503480","display_name":"Face (sociological concept)","level":2,"score":0.4778999984264374},{"id":"https://openalex.org/C2988416141","wikidata":"https://www.wikidata.org/wiki/Q6031139","display_name":"Information loss","level":2,"score":0.47440001368522644},{"id":"https://openalex.org/C2780226545","wikidata":"https://www.wikidata.org/wiki/Q6888030","display_name":"Modality (human\u2013computer interaction)","level":2,"score":0.42590001225471497},{"id":"https://openalex.org/C12725497","wikidata":"https://www.wikidata.org/wiki/Q810247","display_name":"Baseline (sea)","level":2,"score":0.4212999939918518},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.40799999237060547},{"id":"https://openalex.org/C2779903281","wikidata":"https://www.wikidata.org/wiki/Q6888026","display_name":"Modalities","level":2,"score":0.4002000093460083},{"id":"https://openalex.org/C2778112365","wikidata":"https://www.wikidata.org/wiki/Q3511065","display_name":"Sequence (biology)","level":2,"score":0.3993000090122223},{"id":"https://openalex.org/C71139939","wikidata":"https://www.wikidata.org/wiki/Q910194","display_name":"Modal","level":2,"score":0.39800000190734863},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.3544999957084656},{"id":"https://openalex.org/C2780910867","wikidata":"https://www.wikidata.org/wiki/Q1952416","display_name":"Multimodality","level":2,"score":0.3504999876022339},{"id":"https://openalex.org/C111335779","wikidata":"https://www.wikidata.org/wiki/Q3454686","display_name":"Reduction (mathematics)","level":2,"score":0.32339999079704285},{"id":"https://openalex.org/C48677424","wikidata":"https://www.wikidata.org/wiki/Q6888088","display_name":"Mode (computer interface)","level":2,"score":0.30799999833106995},{"id":"https://openalex.org/C774472","wikidata":"https://www.wikidata.org/wiki/Q6760393","display_name":"Margin (machine learning)","level":2,"score":0.30480000376701355},{"id":"https://openalex.org/C36464697","wikidata":"https://www.wikidata.org/wiki/Q451553","display_name":"Visualization","level":2,"score":0.304500013589859},{"id":"https://openalex.org/C52622258","wikidata":"https://www.wikidata.org/wiki/Q131222","display_name":"Information theory","level":2,"score":0.3009999990463257},{"id":"https://openalex.org/C2779954242","wikidata":"https://www.wikidata.org/wiki/Q6031227","display_name":"Information structure","level":2,"score":0.2703000009059906},{"id":"https://openalex.org/C107457646","wikidata":"https://www.wikidata.org/wiki/Q207434","display_name":"Human\u2013computer interaction","level":1,"score":0.266400009393692},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.266400009393692},{"id":"https://openalex.org/C38764148","wikidata":"https://www.wikidata.org/wiki/Q17098245","display_name":"Interaction information","level":2,"score":0.25440001487731934},{"id":"https://openalex.org/C113336015","wikidata":"https://www.wikidata.org/wiki/Q574010","display_name":"Complete information","level":2,"score":0.25049999356269836}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2505.10917","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2505.10917","pdf_url":"https://arxiv.org/pdf/2505.10917","source":{"id":"https://openalex.org/S4393918464","display_name":"ArXiv.org","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2505.10917","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2505.10917","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:oai:arXiv.org:2505.10917","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2505.10917","pdf_url":"https://arxiv.org/pdf/2505.10917","source":{"id":"https://openalex.org/S4393918464","display_name":"ArXiv.org","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"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":{"Current":[0],"multimodal":[1,74],"large":[2],"language":[3],"models":[4,143],"(MLLMs)":[5],"face":[6],"a":[7,15,31,61,86,147],"critical":[8],"challenge":[9],"in":[10,41,162],"modality":[11],"alignment,":[12],"often":[13],"exhibiting":[14],"bias":[16],"towards":[17],"textual":[18],"information":[19,75],"at":[20],"the":[21,36,107,115,157],"expense":[22],"of":[23,35,63,109,119],"other":[24],"modalities":[25],"like":[26],"vision.":[27],"This":[28],"paper":[29],"conducts":[30],"systematic":[32],"information-theoretic":[33],"analysis":[34],"widely":[37],"used":[38],"cross-entropy":[39],"loss":[40],"MLLMs,":[42],"uncovering":[43],"its":[44],"implicit":[45,54],"alignment":[46,65,98,165],"objective.":[47],"Our":[48,138],"theoretical":[49,92],"investigation":[50],"reveals":[51],"that":[52],"this":[53],"objective":[55,99],"has":[56],"inherent":[57],"limitations,":[58],"leading":[59],"to":[60,101],"degradation":[62,108],"cross-modal":[64,103],"as":[66],"text":[67],"sequence":[68],"length":[69],"increases,":[70],"thereby":[71],"hindering":[72],"effective":[73],"fusion.":[76],"To":[77],"overcome":[78],"these":[79],"drawbacks,":[80],"we":[81],"propose":[82],"Vision-Text":[83],"Alignment":[84],"(VISTA),":[85],"novel":[87],"approach":[88],"guided":[89],"by":[90],"our":[91],"insights.":[93],"VISTA":[94,113],"introduces":[95],"an":[96],"explicit":[97],"designed":[100],"maximize":[102],"mutual":[104],"information,":[105],"preventing":[106],"visual":[110,116],"alignment.":[111],"Notably,":[112],"enhances":[114],"understanding":[117],"capabilities":[118],"existing":[120],"MLLMs":[121],"without":[122],"requiring":[123],"any":[124],"additional":[125],"trainable":[126],"modules":[127],"or":[128],"extra":[129],"training":[130],"data,":[131],"making":[132],"it":[133],"both":[134],"efficient":[135],"and":[136,154],"practical.":[137],"method":[139],"significantly":[140],"outperforms":[141],"baseline":[142],"across":[144],"more":[145],"than":[146],"dozen":[148],"benchmark":[149],"datasets,":[150],"including":[151],"VQAv2,":[152],"MMStar,":[153],"MME,":[155],"paving":[156],"way":[158],"for":[159],"new":[160],"directions":[161],"MLLM":[163],"modal":[164],"research.":[166]},"counts_by_year":[],"updated_date":"2026-04-21T08:09:41.155169","created_date":"2025-10-10T00:00:00"}