{"id":"https://openalex.org/W7157874233","doi":"https://doi.org/10.48550/arxiv.2604.25273","title":"Combating Visual Neglect and Semantic Drift in Large Multimodal Models for Enhanced Cross-Modal Retrieval","display_name":"Combating Visual Neglect and Semantic Drift in Large Multimodal Models for Enhanced Cross-Modal Retrieval","publication_year":2026,"publication_date":"2026-04-28","ids":{"openalex":"https://openalex.org/W7157874233","doi":"https://doi.org/10.48550/arxiv.2604.25273"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2604.25273","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.25273","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":"Preprint"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2604.25273","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5101544856","display_name":"Guosheng Zhang","orcid":"https://orcid.org/0000-0003-4041-7911"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Guosheng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134829905","display_name":"Linkai Liu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Linkai","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5121607946","display_name":"Keyao Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Keyao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5042201593","display_name":"Haixiao Yue","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yue, Haixiao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134824067","display_name":"Zhiwen Tan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tan, Zhiwen","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5134834658","display_name":"Xiao Tan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tan, Xiao","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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.7355999946594238,"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":0.7355999946594238,"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/T11605","display_name":"Visual Attention and Saliency Detection","score":0.21539999544620514,"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":0.011900000274181366,"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/interpretability","display_name":"Interpretability","score":0.8047999739646912},{"id":"https://openalex.org/keywords/salient","display_name":"Salient","score":0.7628999948501587},{"id":"https://openalex.org/keywords/embedding","display_name":"Embedding","score":0.5968000292778015},{"id":"https://openalex.org/keywords/feature","display_name":"Feature (linguistics)","score":0.5730999708175659},{"id":"https://openalex.org/keywords/visualization","display_name":"Visualization","score":0.47510001063346863},{"id":"https://openalex.org/keywords/modality","display_name":"Modality (human\u2013computer interaction)","score":0.46810001134872437},{"id":"https://openalex.org/keywords/neglect","display_name":"Neglect","score":0.4350999891757965},{"id":"https://openalex.org/keywords/representation","display_name":"Representation (politics)","score":0.43130001425743103},{"id":"https://openalex.org/keywords/focus","display_name":"Focus (optics)","score":0.40310001373291016}],"concepts":[{"id":"https://openalex.org/C2781067378","wikidata":"https://www.wikidata.org/wiki/Q17027399","display_name":"Interpretability","level":2,"score":0.8047999739646912},{"id":"https://openalex.org/C2780719617","wikidata":"https://www.wikidata.org/wiki/Q1030752","display_name":"Salient","level":2,"score":0.7628999948501587},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7430999875068665},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.6837000250816345},{"id":"https://openalex.org/C41608201","wikidata":"https://www.wikidata.org/wiki/Q980509","display_name":"Embedding","level":2,"score":0.5968000292778015},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.5730999708175659},{"id":"https://openalex.org/C36464697","wikidata":"https://www.wikidata.org/wiki/Q451553","display_name":"Visualization","level":2,"score":0.47510001063346863},{"id":"https://openalex.org/C2780226545","wikidata":"https://www.wikidata.org/wiki/Q6888030","display_name":"Modality (human\u2013computer interaction)","level":2,"score":0.46810001134872437},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.46549999713897705},{"id":"https://openalex.org/C2776289891","wikidata":"https://www.wikidata.org/wiki/Q1931511","display_name":"Neglect","level":2,"score":0.4350999891757965},{"id":"https://openalex.org/C2776359362","wikidata":"https://www.wikidata.org/wiki/Q2145286","display_name":"Representation (politics)","level":3,"score":0.43130001425743103},{"id":"https://openalex.org/C192209626","wikidata":"https://www.wikidata.org/wiki/Q190909","display_name":"Focus (optics)","level":2,"score":0.40310001373291016},{"id":"https://openalex.org/C2781122975","wikidata":"https://www.wikidata.org/wiki/Q16928266","display_name":"Semantic feature","level":2,"score":0.3959999978542328},{"id":"https://openalex.org/C2780910867","wikidata":"https://www.wikidata.org/wiki/Q1952416","display_name":"Multimodality","level":2,"score":0.3874000012874603},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.38029998540878296},{"id":"https://openalex.org/C184337299","wikidata":"https://www.wikidata.org/wiki/Q1437428","display_name":"Semantics (computer science)","level":2,"score":0.3287000060081482},{"id":"https://openalex.org/C107457646","wikidata":"https://www.wikidata.org/wiki/Q207434","display_name":"Human\u2013computer interaction","level":1,"score":0.3098999857902527},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.27639999985694885},{"id":"https://openalex.org/C2780660688","wikidata":"https://www.wikidata.org/wiki/Q25052564","display_name":"Multimodal learning","level":2,"score":0.27300000190734863},{"id":"https://openalex.org/C55439883","wikidata":"https://www.wikidata.org/wiki/Q360812","display_name":"Correctness","level":2,"score":0.2709999978542328},{"id":"https://openalex.org/C59404180","wikidata":"https://www.wikidata.org/wiki/Q17013334","display_name":"Feature learning","level":2,"score":0.26660001277923584},{"id":"https://openalex.org/C165696696","wikidata":"https://www.wikidata.org/wiki/Q11287","display_name":"Exploit","level":2,"score":0.26570001244544983},{"id":"https://openalex.org/C66024118","wikidata":"https://www.wikidata.org/wiki/Q1122506","display_name":"Computational model","level":2,"score":0.2606000006198883},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.25519999861717224},{"id":"https://openalex.org/C87833898","wikidata":"https://www.wikidata.org/wiki/Q1060280","display_name":"Advanced driver assistance systems","level":2,"score":0.2531999945640564},{"id":"https://openalex.org/C178253425","wikidata":"https://www.wikidata.org/wiki/Q162668","display_name":"Visual perception","level":3,"score":0.2531000077724457}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2604.25273","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.25273","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":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2604.25273","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.25273","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":"Preprint"},"sustainable_development_goals":[{"display_name":"Quality Education","score":0.7071078419685364,"id":"https://metadata.un.org/sdg/4"}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Despite":[0],"significant":[1],"progress":[2],"in":[3,42,59,79,124],"Unified":[4],"Multimodal":[5,11,96],"Retrieval":[6],"(UMR)":[7],"powered":[8],"by":[9],"Large":[10],"Models":[12],"(LMMs),":[13],"existing":[14],"embedding":[15],"methods":[16],"primarily":[17],"focus":[18],"on":[19,75,150,173],"sample-level":[20],"objectives":[21],"via":[22],"contrastive":[23],"learning":[24,107],"while":[25],"overlooking":[26],"the":[27,34,151,174,191],"crucial":[28],"subject-level":[29,180],"semantics.":[30],"This":[31],"limitation":[32],"hinders":[33],"model's":[35],"ability":[36],"to":[37,53,66,73,103,117,132],"group":[38],"semantically":[39,138,159],"coherent":[40,160],"subjects":[41],"complex":[43],"multimodal":[44,184],"queries,":[45],"manifesting":[46],"as":[47],"semantic":[48],"alignment":[49],"deviation--where":[50],"models":[51],"fail":[52],"accurately":[54],"localize":[55],"salient":[56,68,121],"text-referred":[57],"regions":[58],"visual":[60,69,80,87,115,122,147],"content.":[61],"Moreover,":[62],"without":[63],"explicit":[64],"guidance":[65],"model":[67],"subjects,":[70],"LMMs":[71,113],"tend":[72],"over-rely":[74],"textual":[76],"cues,":[77],"resulting":[78],"modality":[81],"neglect":[82],"and":[83,114,119,127,158,193],"suboptimal":[84],"utilization":[85],"of":[86,195],"knowledge.":[88],"To":[89],"this":[90],"end,":[91],"we":[92],"propose":[93],"Salient":[94],"Subject-Aware":[95],"Embedding":[97],"(SSA-ME),":[98],"a":[99,129,142,156],"novel":[100],"framework":[101],"designed":[102],"enhance":[104],"fine-grained":[105],"representation":[106],"through":[108],"saliency-aware":[109],"modeling.":[110],"SSA-ME":[111],"leverages":[112],"experts":[116],"identify":[118],"emphasize":[120],"concepts":[123],"image-text":[125],"pairs,":[126],"introduces":[128],"saliency-guided":[130],"objective":[131],"better":[133],"align":[134],"cross-modal":[135],"attention":[136],"with":[137],"meaningful":[139],"regions.":[140],"Additionally,":[141],"feature":[143],"regeneration":[144],"module":[145],"recalibrates":[146],"features":[148],"based":[149],"derived":[152],"saliency":[153],"maps,":[154],"ensuring":[155],"balanced":[157],"integration":[161],"across":[162],"modalities.":[163],"Extensive":[164],"experiments":[165],"show":[166],"that":[167,178],"our":[168,196],"method":[169],"achieves":[170],"state-of-the-art":[171],"performance":[172],"MMEB":[175],"benchmark,":[176],"demonstrating":[177],"incorporating":[179],"modeling":[181],"substantially":[182],"improves":[183],"retrieval.":[185],"Comprehensive":[186],"qualitative":[187],"analyses":[188],"further":[189],"illustrate":[190],"interpretability":[192],"effectiveness":[194],"approach.":[197]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-04-30T00:00:00"}