{"id":"https://openalex.org/W4409050808","doi":"https://doi.org/10.1109/iccv51701.2025.02205","title":"LangBridge: Interpreting Image as a Combination of Language Embeddings","display_name":"LangBridge: Interpreting Image as a Combination of Language Embeddings","publication_year":2025,"publication_date":"2025-10-19","ids":{"openalex":"https://openalex.org/W4409050808","doi":"https://doi.org/10.1109/iccv51701.2025.02205"},"language":"en","primary_location":{"id":"doi:10.1109/iccv51701.2025.02205","is_oa":false,"landing_page_url":"https://doi.org/10.1109/iccv51701.2025.02205","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE/CVF International Conference on Computer Vision (ICCV)","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["arxiv","crossref","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2503.19404","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5061619815","display_name":"Jiaqi Liao","orcid":null},"institutions":[{"id":"https://openalex.org/I4391012619","display_name":"Shanghai Artificial Intelligence Laboratory","ror":"https://ror.org/03wkvpx79","country_code":null,"type":"facility","lineage":["https://openalex.org/I4391012619"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jiaqi Liao","raw_affiliation_strings":["Shanghai AI Laboratory,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Shanghai AI Laboratory,China","institution_ids":["https://openalex.org/I4391012619"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101367498","display_name":"Yuwei Niu","orcid":null},"institutions":[{"id":"https://openalex.org/I20231570","display_name":"Peking University","ror":"https://ror.org/02v51f717","country_code":"CN","type":"education","lineage":["https://openalex.org/I20231570"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yuwei Niu","raw_affiliation_strings":["Peking University,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Peking University,China","institution_ids":["https://openalex.org/I20231570"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5104225369","display_name":"Fanqing Meng","orcid":null},"institutions":[{"id":"https://openalex.org/I4391012619","display_name":"Shanghai Artificial Intelligence Laboratory","ror":"https://ror.org/03wkvpx79","country_code":null,"type":"facility","lineage":["https://openalex.org/I4391012619"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Fanqing Meng","raw_affiliation_strings":["Shanghai AI Laboratory,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Shanghai AI Laboratory,China","institution_ids":["https://openalex.org/I4391012619"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100348619","display_name":"Hao Li","orcid":"https://orcid.org/0000-0002-6336-8772"},"institutions":[{"id":"https://openalex.org/I4391012619","display_name":"Shanghai Artificial Intelligence Laboratory","ror":"https://ror.org/03wkvpx79","country_code":null,"type":"facility","lineage":["https://openalex.org/I4391012619"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Hao Li","raw_affiliation_strings":["Shanghai AI Laboratory,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Shanghai AI Laboratory,China","institution_ids":["https://openalex.org/I4391012619"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5087792832","display_name":"Changyao Tian","orcid":"https://orcid.org/0000-0002-3285-4671"},"institutions":[{"id":"https://openalex.org/I4391012619","display_name":"Shanghai Artificial Intelligence Laboratory","ror":"https://ror.org/03wkvpx79","country_code":null,"type":"facility","lineage":["https://openalex.org/I4391012619"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Changyao Tian","raw_affiliation_strings":["Shanghai AI Laboratory,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Shanghai AI Laboratory,China","institution_ids":["https://openalex.org/I4391012619"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5065179266","display_name":"Yinuo Du","orcid":"https://orcid.org/0009-0000-9127-4063"},"institutions":[{"id":"https://openalex.org/I4391012619","display_name":"Shanghai Artificial Intelligence Laboratory","ror":"https://ror.org/03wkvpx79","country_code":null,"type":"facility","lineage":["https://openalex.org/I4391012619"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yinuo Du Yuwen Xiong","raw_affiliation_strings":["Shanghai AI Laboratory,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Shanghai AI Laboratory,China","institution_ids":["https://openalex.org/I4391012619"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5077530975","display_name":"Dianqi Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Dianqi Li","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5112877229","display_name":"Xizhou Zhu","orcid":null},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xizhou Zhu","raw_affiliation_strings":["Tsinghua University,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Tsinghua University,China","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Li Yuan","orcid":null},"institutions":[{"id":"https://openalex.org/I20231570","display_name":"Peking University","ror":"https://ror.org/02v51f717","country_code":"CN","type":"education","lineage":["https://openalex.org/I20231570"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Li Yuan","raw_affiliation_strings":["Peking University,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Peking University,China","institution_ids":["https://openalex.org/I20231570"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Jifeng Dai","orcid":null},"institutions":[{"id":"https://openalex.org/I4391012619","display_name":"Shanghai Artificial Intelligence Laboratory","ror":"https://ror.org/03wkvpx79","country_code":null,"type":"facility","lineage":["https://openalex.org/I4391012619"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jifeng Dai","raw_affiliation_strings":["Shanghai AI Laboratory,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Shanghai AI Laboratory,China","institution_ids":["https://openalex.org/I4391012619"]}]},{"author_position":"last","author":{"id":null,"display_name":"Yu Cheng","orcid":null},"institutions":[{"id":"https://openalex.org/I177725633","display_name":"Chinese University of Hong Kong","ror":"https://ror.org/00t33hh48","country_code":"HK","type":"education","lineage":["https://openalex.org/I177725633"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Yu Cheng","raw_affiliation_strings":["The Chinese University of Hong Kong,Hong Kong"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Chinese University of Hong Kong,Hong Kong","institution_ids":["https://openalex.org/I177725633"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":4,"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":"23752","last_page":"23762"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T12151","display_name":"Interpreting and Communication in Healthcare","score":0.9787999987602234,"subfield":{"id":"https://openalex.org/subfields/3600","display_name":"General Health Professions"},"field":{"id":"https://openalex.org/fields/36","display_name":"Health Professions"},"domain":{"id":"https://openalex.org/domains/4","display_name":"Health Sciences"}},"topics":[{"id":"https://openalex.org/T12151","display_name":"Interpreting and Communication in Healthcare","score":0.9787999987602234,"subfield":{"id":"https://openalex.org/subfields/3600","display_name":"General Health Professions"},"field":{"id":"https://openalex.org/fields/36","display_name":"Health Professions"},"domain":{"id":"https://openalex.org/domains/4","display_name":"Health Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/image","display_name":"Image (mathematics)","score":0.6315694451332092},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.46871986985206604},{"id":"https://openalex.org/keywords/natural-language-processing","display_name":"Natural language processing","score":0.45454901456832886},{"id":"https://openalex.org/keywords/linguistics","display_name":"Linguistics","score":0.40127670764923096},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.35986578464508057},{"id":"https://openalex.org/keywords/philosophy","display_name":"Philosophy","score":0.19411975145339966}],"concepts":[{"id":"https://openalex.org/C115961682","wikidata":"https://www.wikidata.org/wiki/Q860623","display_name":"Image (mathematics)","level":2,"score":0.6315694451332092},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.46871986985206604},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.45454901456832886},{"id":"https://openalex.org/C41895202","wikidata":"https://www.wikidata.org/wiki/Q8162","display_name":"Linguistics","level":1,"score":0.40127670764923096},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.35986578464508057},{"id":"https://openalex.org/C138885662","wikidata":"https://www.wikidata.org/wiki/Q5891","display_name":"Philosophy","level":0,"score":0.19411975145339966}],"mesh":[],"locations_count":3,"locations":[{"id":"doi:10.1109/iccv51701.2025.02205","is_oa":false,"landing_page_url":"https://doi.org/10.1109/iccv51701.2025.02205","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE/CVF International Conference on Computer Vision (ICCV)","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2503.19404","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2503.19404","pdf_url":"https://arxiv.org/pdf/2503.19404","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2503.19404","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2503.19404","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":"pmh:oai:arXiv.org:2503.19404","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2503.19404","pdf_url":"https://arxiv.org/pdf/2503.19404","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":"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":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W2390279801","https://openalex.org/W4391913857","https://openalex.org/W2358668433","https://openalex.org/W4396701345","https://openalex.org/W2376932109","https://openalex.org/W2001405890","https://openalex.org/W3204019825"],"abstract_inverted_index":{"Recent":[0],"years":[1],"have":[2,12,79],"witnessed":[3],"remarkable":[4],"advances":[5],"in":[6,197],"Large":[7],"Vision-Language":[8],"Models":[9],"(LVLMs),":[10],"which":[11],"achieved":[13],"human-level":[14],"performance":[15,214],"across":[16,154,208],"various":[17],"complex":[18],"vision-language":[19,191],"tasks.":[20],"Following":[21],"LLaVA's":[22],"paradigm,":[23],"mainstream":[24],"LVLMs":[25],"typically":[26],"employ":[27],"a":[28,35,132,165],"shallow":[29],"MLP":[30,87,106],"for":[31,40],"visual-language":[32],"alignment":[33,42,83,192],"through":[34],"two-stage":[36],"training":[37],"process:":[38],"pretraining":[39],"cross-modal":[41],"followed":[43],"by":[44,120,193],"instruction":[45],"tuning.":[46],"While":[47],"this":[48,127],"approach":[49],"has":[50,69],"proven":[51],"effective,":[52],"the":[53,60,81,86,102],"underlying":[54],"mechanisms":[55],"of":[56,105,143],"how":[57,71],"MLPs":[58],"bridge":[59],"modality":[61],"gap":[62],"remain":[63],"poorly":[64],"understood.":[65],"Although":[66],"some":[67],"research":[68],"explored":[70],"LLMs":[72,156,210],"process":[73],"transformed":[74],"visual":[75,115,138,195],"tokens,":[76],"few":[77],"studies":[78],"investigated":[80],"fundamental":[82],"mechanism.":[84],"Furthermore,":[85],"adapter":[88,134,152,167],"requires":[89],"retraining":[90],"whenever":[91],"switching":[92],"LLM":[93,144,198],"backbones.":[94],"To":[95],"address":[96],"these":[97],"limitations,":[98],"we":[99,129],"first":[100],"investigate":[101],"working":[103],"principles":[104],"adapters":[107],"and":[108],"discover":[109],"that":[110,135,164],"they":[111],"learn":[112],"to":[113,140,175],"project":[114,218],"embeddings":[116,123],"into":[117],"subspaces":[118],"spanned":[119],"corresponding":[121],"text":[122],"progressively.":[124],"Based":[125],"on":[126,169],"insight,":[128],"propose":[130],"LangBridge,":[131],"novel":[133],"explicitly":[136],"maps":[137],"tokens":[139],"linear":[141],"combinations":[142],"vocabulary":[145],"embeddings.":[146],"This":[147],"innovative":[148],"design":[149,204],"enables":[150,189],"pretraining-free":[151],"transfer":[153],"different":[155],"while":[157,183,201],"maintaining":[158,184],"performance.":[159,186],"Our":[160],"experimental":[161],"results":[162],"demonstrate":[163],"LangBridge":[166,188],"pre-trained":[168],"Qwen2-0.5B":[170],"can":[171],"be":[172],"directly":[173],"applied":[174],"larger":[176],"models":[177],"such":[178],"as":[179],"LLaMA3-8B":[180],"or":[181],"Qwen2.5-14B":[182],"competitive":[185],"Overall,":[187],"interpretable":[190],"grounding":[194],"representations":[196],"vocab":[199],"embedding,":[200],"its":[202],"plug-and-play":[203],"ensures":[205],"efficient":[206],"reuse":[207],"multiple":[209],"with":[211],"nearly":[212],"no":[213],"degradation.":[215],"See":[216],"our":[217],"page":[219],"at":[220],"https://curryx-001.github.io/LangBridge.github.io/":[221]},"counts_by_year":[],"updated_date":"2026-07-21T08:15:58.654021","created_date":"2025-10-10T00:00:00"}