{"id":"https://openalex.org/W4414359230","doi":"https://doi.org/10.24963/ijcai.2025/678","title":"MsRAG: Knowledge Augumented Image Captioning with Object-level Multi-source RAG","display_name":"MsRAG: Knowledge Augumented Image Captioning with Object-level Multi-source RAG","publication_year":2025,"publication_date":"2025-09-01","ids":{"openalex":"https://openalex.org/W4414359230","doi":"https://doi.org/10.24963/ijcai.2025/678"},"language":"en","primary_location":{"id":"doi:10.24963/ijcai.2025/678","is_oa":false,"landing_page_url":"https://doi.org/10.24963/ijcai.2025/678","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":false,"oa_status":"closed","oa_url":null,"any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5113062347","display_name":"Yuming Qiao","orcid":null},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Yuming Qiao","raw_affiliation_strings":["OPPO Research Institute"],"affiliations":[{"raw_affiliation_string":"OPPO Research Institute","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102902110","display_name":"Yuechen Wang","orcid":"https://orcid.org/0000-0001-6098-6164"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yuechen Wang","raw_affiliation_strings":["OPPO Research Institute"],"affiliations":[{"raw_affiliation_string":"OPPO Research Institute","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5064432169","display_name":"Dan Meng","orcid":"https://orcid.org/0000-0003-0003-0659"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Dan Meng","raw_affiliation_strings":["OPPO Research Institute"],"affiliations":[{"raw_affiliation_string":"OPPO Research Institute","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5007720098","display_name":"H. Lu","orcid":"https://orcid.org/0000-0001-6332-2785"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Haonan Lu","raw_affiliation_strings":["OPPO AI Center"],"affiliations":[{"raw_affiliation_string":"OPPO AI Center","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100433323","display_name":"Zhenyu Yang","orcid":"https://orcid.org/0000-0002-0773-0298"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhenyu Yang","raw_affiliation_strings":["OPPO AI Center"],"affiliations":[{"raw_affiliation_string":"OPPO AI Center","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5100441859","display_name":"Xudong Zhang","orcid":"https://orcid.org/0000-0002-9666-405X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xudong Zhang","raw_affiliation_strings":["OPPO Research Institute"],"affiliations":[{"raw_affiliation_string":"OPPO Research Institute","institution_ids":[]}]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":6,"corresponding_author_ids":["https://openalex.org/A5113062347"],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.26670807,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"6093","last_page":"6101"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11714","display_name":"Multimodal Machine Learning Applications","score":0.9997000098228455,"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":0.9997000098228455,"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/T11439","display_name":"Video Analysis and Summarization","score":0.9549999833106995,"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/T10627","display_name":"Advanced Image and Video Retrieval Techniques","score":0.9337999820709229,"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/closed-captioning","display_name":"Closed captioning","score":0.9758999943733215},{"id":"https://openalex.org/keywords/leverage","display_name":"Leverage (statistics)","score":0.732699990272522},{"id":"https://openalex.org/keywords/robustness","display_name":"Robustness (evolution)","score":0.6276000142097473},{"id":"https://openalex.org/keywords/visualization","display_name":"Visualization","score":0.46880000829696655},{"id":"https://openalex.org/keywords/limiting","display_name":"Limiting","score":0.40959998965263367},{"id":"https://openalex.org/keywords/scope","display_name":"Scope (computer science)","score":0.37040001153945923},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.3605000078678131}],"concepts":[{"id":"https://openalex.org/C157657479","wikidata":"https://www.wikidata.org/wiki/Q2367247","display_name":"Closed captioning","level":3,"score":0.9758999943733215},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8374000191688538},{"id":"https://openalex.org/C153083717","wikidata":"https://www.wikidata.org/wiki/Q6535263","display_name":"Leverage (statistics)","level":2,"score":0.732699990272522},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.6276000142097473},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.571399986743927},{"id":"https://openalex.org/C36464697","wikidata":"https://www.wikidata.org/wiki/Q451553","display_name":"Visualization","level":2,"score":0.46880000829696655},{"id":"https://openalex.org/C188198153","wikidata":"https://www.wikidata.org/wiki/Q1613840","display_name":"Limiting","level":2,"score":0.40959998965263367},{"id":"https://openalex.org/C2778012447","wikidata":"https://www.wikidata.org/wiki/Q1034415","display_name":"Scope (computer science)","level":2,"score":0.37040001153945923},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.3605000078678131},{"id":"https://openalex.org/C100776233","wikidata":"https://www.wikidata.org/wiki/Q2532492","display_name":"Bridge (graph theory)","level":2,"score":0.35910001397132874},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.3578999936580658},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3490000069141388},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.33160001039505005},{"id":"https://openalex.org/C2777508537","wikidata":"https://www.wikidata.org/wiki/Q7936620","display_name":"Visual reasoning","level":2,"score":0.3190000057220459},{"id":"https://openalex.org/C86034646","wikidata":"https://www.wikidata.org/wiki/Q474311","display_name":"Semantic gap","level":4,"score":0.3149999976158142},{"id":"https://openalex.org/C23123220","wikidata":"https://www.wikidata.org/wiki/Q816826","display_name":"Information retrieval","level":1,"score":0.31360000371932983},{"id":"https://openalex.org/C2780878386","wikidata":"https://www.wikidata.org/wiki/Q1659648","display_name":"Visual language","level":2,"score":0.3109000027179718},{"id":"https://openalex.org/C175154964","wikidata":"https://www.wikidata.org/wiki/Q380077","display_name":"Task analysis","level":3,"score":0.30059999227523804},{"id":"https://openalex.org/C107457646","wikidata":"https://www.wikidata.org/wiki/Q207434","display_name":"Human\u2013computer interaction","level":1,"score":0.2948000133037567},{"id":"https://openalex.org/C158495155","wikidata":"https://www.wikidata.org/wiki/Q2369151","display_name":"Visual search","level":2,"score":0.2736999988555908},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.2558000087738037},{"id":"https://openalex.org/C172367668","wikidata":"https://www.wikidata.org/wiki/Q6504956","display_name":"Data visualization","level":3,"score":0.25380000472068787}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.24963/ijcai.2025/678","is_oa":false,"landing_page_url":"https://doi.org/10.24963/ijcai.2025/678","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Language-Visual":[0],"Large":[1],"Models":[2],"(LVLMs)":[3],"have":[4],"made":[5],"significant":[6],"strides":[7],"in":[8,25,51,78],"enhancing":[9,115],"visual":[10,20,108,157,171,177],"understanding":[11],"capabilities.":[12],"However,":[13],"these":[14,59],"models":[15],"often":[16],"struggle":[17],"with":[18,179],"knowledge-based":[19],"tasks":[21],"due":[22],"to":[23,69,141,160],"constrains":[24],"their":[26,49,180],"pre-training":[27],"data":[28,144],"scope":[29],"and":[30,72,84,111,118,169,196],"timeliness.":[31],"Existing":[32],"Retrieval-Augmented":[33],"Generation":[34],"(RAG)":[35],"methods":[36],"can":[37],"effectively":[38,70],"solve":[39],"the":[40,79,116,162,186,203],"problem":[41],"but":[42],"primarily":[43],"rely":[44],"on":[45,135],"user":[46,82],"queries,":[47,83],"limiting":[48],"applicability":[50],"scenarios":[52],"without":[53],"explicit":[54],"language":[55],"input.":[56],"To":[57,184],"overcome":[58],"challenges,":[60],"we":[61,190],"introduce":[62],"MsRAG,":[63,189],"a":[64,155,192],"knowledge-augmented":[65],"captioning":[66,87],"framework":[67],"designed":[68],"retrieve":[71],"utilize":[73],"external":[74],"real-world":[75],"knowledge,":[76],"particularly":[77],"absence":[80],"of":[81,120,176,188,194,205],"perform":[85],"dense":[86],"for":[88],"subjects.":[89],"MsRAG":[90,206],"comprises":[91],"three":[92],"key":[93],"components:":[94],"(1)":[95],"Parallel":[96],"Visual":[97],"Search":[98],"Module.":[99],"It":[100],"retrieves":[101],"fine-grained":[102],"object-level":[103],"knowledge":[104],"using":[105],"both":[106],"online":[107],"search":[109],"engines":[110],"offline":[112],"domain-knowledge":[113],"databases,":[114],"robustness":[117],"richness":[119],"retrieved":[121,136],"information.":[122],"(2)":[123],"Prompt":[124],"Templates":[125],"Pool.":[126],"The":[127,199],"prompt":[128],"pool":[129],"dynamically":[130],"assigns":[131],"appropriate":[132],"prompts":[133],"based":[134],"information,":[137],"optimizing":[138],"LVLMs'":[139],"ability":[140],"leverage":[142],"relevant":[143],"under":[145],"complex":[146],"RAG":[147,167,182],"conditions.":[148],"(3)":[149],"Visual-RAG":[150],"Alignment":[151],"Module,":[152],"which":[153],"employs":[154],"novel":[156],"prompting":[158],"method":[159],"bridge":[161],"modality":[163],"gap":[164],"between":[165],"textual":[166],"content":[168],"corresponding":[170],"objects,":[172],"enabling":[173],"precise":[174],"alignment":[175],"elements":[178],"text-format":[181],"content.":[183],"validate":[185],"effectiveness":[187],"conducted":[191],"series":[193],"qualitative":[195],"quantitative":[197],"experiments.":[198],"evaluation":[200],"results":[201],"demonstrate":[202],"superiority":[204],"over":[207],"other":[208],"methods.":[209]},"counts_by_year":[],"updated_date":"2026-03-07T16:01:11.037858","created_date":"2025-10-10T00:00:00"}