{"id":"https://openalex.org/W4411635651","doi":"https://doi.org/10.1145/3731715.3733314","title":"DomainDiff: Unified Two-Stage Optimization for Text-Video Retrieval","display_name":"DomainDiff: Unified Two-Stage Optimization for Text-Video Retrieval","publication_year":2025,"publication_date":"2025-06-25","ids":{"openalex":"https://openalex.org/W4411635651","doi":"https://doi.org/10.1145/3731715.3733314"},"language":"en","primary_location":{"id":"doi:10.1145/3731715.3733314","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3731715.3733314","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2025 International Conference on Multimedia Retrieval","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/A5002181176","display_name":"C. Wang","orcid":null},"institutions":[{"id":"https://openalex.org/I186272606","display_name":"Guangdong University of Foreign Studies","ror":"https://ror.org/00fhc9y79","country_code":"CN","type":"education","lineage":["https://openalex.org/I186272606"]}],"countries":["CN"],"is_corresponding":true,"raw_author_name":"Chenxu Wang","raw_affiliation_strings":["Guangdong University of Foreign Studies, Guangzhou, China"],"raw_orcid":"https://orcid.org/0009-0004-2031-7635","affiliations":[{"raw_affiliation_string":"Guangdong University of Foreign Studies, Guangzhou, China","institution_ids":["https://openalex.org/I186272606"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100639988","display_name":"Dong Zhou","orcid":"https://orcid.org/0000-0002-3310-8347"},"institutions":[{"id":"https://openalex.org/I186272606","display_name":"Guangdong University of Foreign Studies","ror":"https://ror.org/00fhc9y79","country_code":"CN","type":"education","lineage":["https://openalex.org/I186272606"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Dong Zhou","raw_affiliation_strings":["Guangdong University of Foreign Studies, Guangzhou, China"],"raw_orcid":"https://orcid.org/0000-0002-3310-8347","affiliations":[{"raw_affiliation_string":"Guangdong University of Foreign Studies, Guangzhou, China","institution_ids":["https://openalex.org/I186272606"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5010055657","display_name":"Jianghao Lin","orcid":"https://orcid.org/0000-0002-5935-2110"},"institutions":[{"id":"https://openalex.org/I186272606","display_name":"Guangdong University of Foreign Studies","ror":"https://ror.org/00fhc9y79","country_code":"CN","type":"education","lineage":["https://openalex.org/I186272606"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jianghao Lin","raw_affiliation_strings":["Guangdong University of Foreign Studies, Guangzhou, China"],"raw_orcid":"https://orcid.org/0000-0002-5935-2110","affiliations":[{"raw_affiliation_string":"Guangdong University of Foreign Studies, Guangzhou, China","institution_ids":["https://openalex.org/I186272606"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5115076610","display_name":"Yongmei Zhou","orcid":"https://orcid.org/0000-0003-2661-3078"},"institutions":[{"id":"https://openalex.org/I186272606","display_name":"Guangdong University of Foreign Studies","ror":"https://ror.org/00fhc9y79","country_code":"CN","type":"education","lineage":["https://openalex.org/I186272606"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yongmei Zhou","raw_affiliation_strings":["Guangdong University of Foreign Studies, Guangzhou, China"],"raw_orcid":"https://orcid.org/0000-0003-2661-3078","affiliations":[{"raw_affiliation_string":"Guangdong University of Foreign Studies, Guangzhou, China","institution_ids":["https://openalex.org/I186272606"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5101732032","display_name":"Aimin Yang","orcid":"https://orcid.org/0009-0002-1751-4801"},"institutions":[{"id":"https://openalex.org/I154833797","display_name":"Lingnan Normal University","ror":"https://ror.org/01h6ecw13","country_code":"CN","type":"education","lineage":["https://openalex.org/I154833797"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Aimin Yang","raw_affiliation_strings":["Lingnan Normal University, Zhanjiang, China"],"raw_orcid":"https://orcid.org/0009-0002-1751-4801","affiliations":[{"raw_affiliation_string":"Lingnan Normal University, Zhanjiang, China","institution_ids":["https://openalex.org/I154833797"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":5,"corresponding_author_ids":["https://openalex.org/A5002181176"],"corresponding_institution_ids":["https://openalex.org/I186272606"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.12147661,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"1432","last_page":"1441"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11714","display_name":"Multimodal Machine Learning Applications","score":0.9998999834060669,"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":0.9998999834060669,"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":0.9997000098228455,"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/T10627","display_name":"Advanced Image and Video Retrieval Techniques","score":0.9993000030517578,"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/computer-science","display_name":"Computer science","score":0.7280926704406738},{"id":"https://openalex.org/keywords/stage","display_name":"Stage (stratigraphy)","score":0.6095255017280579},{"id":"https://openalex.org/keywords/information-retrieval","display_name":"Information retrieval","score":0.4860580265522003},{"id":"https://openalex.org/keywords/video-retrieval","display_name":"Video retrieval","score":0.425533652305603},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.3884657621383667}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7280926704406738},{"id":"https://openalex.org/C146357865","wikidata":"https://www.wikidata.org/wiki/Q1123245","display_name":"Stage (stratigraphy)","level":2,"score":0.6095255017280579},{"id":"https://openalex.org/C23123220","wikidata":"https://www.wikidata.org/wiki/Q816826","display_name":"Information retrieval","level":1,"score":0.4860580265522003},{"id":"https://openalex.org/C2983174267","wikidata":"https://www.wikidata.org/wiki/Q3775098","display_name":"Video retrieval","level":2,"score":0.425533652305603},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3884657621383667},{"id":"https://openalex.org/C86803240","wikidata":"https://www.wikidata.org/wiki/Q420","display_name":"Biology","level":0,"score":0.0},{"id":"https://openalex.org/C151730666","wikidata":"https://www.wikidata.org/wiki/Q7205","display_name":"Paleontology","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3731715.3733314","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3731715.3733314","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2025 International Conference on Multimedia Retrieval","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G7042713332","display_name":null,"funder_award_id":"62376062","funder_id":"https://openalex.org/F4320323817","funder_display_name":"Universitas Brawijaya"},{"id":"https://openalex.org/G8201270149","display_name":null,"funder_award_id":"24BXW047","funder_id":"https://openalex.org/F4320323817","funder_display_name":"Universitas Brawijaya"}],"funders":[{"id":"https://openalex.org/F4320323817","display_name":"Universitas Brawijaya","ror":"https://ror.org/01wk3d929"}],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":50,"referenced_works":["https://openalex.org/W2078238240","https://openalex.org/W2425121537","https://openalex.org/W2525579820","https://openalex.org/W2531409750","https://openalex.org/W2612690371","https://openalex.org/W2963017553","https://openalex.org/W2963916161","https://openalex.org/W2990138404","https://openalex.org/W3035265375","https://openalex.org/W3035356601","https://openalex.org/W3043840704","https://openalex.org/W3154596443","https://openalex.org/W3157506437","https://openalex.org/W3162926177","https://openalex.org/W3174873881","https://openalex.org/W3175939205","https://openalex.org/W3204588463","https://openalex.org/W3204647170","https://openalex.org/W4225384855","https://openalex.org/W4225414521","https://openalex.org/W4247924304","https://openalex.org/W4249502209","https://openalex.org/W4249914127","https://openalex.org/W4255556797","https://openalex.org/W4285606530","https://openalex.org/W4297947725","https://openalex.org/W4303449836","https://openalex.org/W4304014690","https://openalex.org/W4309804076","https://openalex.org/W4312299780","https://openalex.org/W4312661097","https://openalex.org/W4312810897","https://openalex.org/W4312999114","https://openalex.org/W4313156423","https://openalex.org/W4383604991","https://openalex.org/W4386076265","https://openalex.org/W4390872010","https://openalex.org/W4390872366","https://openalex.org/W4390872434","https://openalex.org/W4390873165","https://openalex.org/W4393161221","https://openalex.org/W4395481535","https://openalex.org/W4402704596","https://openalex.org/W4402727787","https://openalex.org/W4403511263","https://openalex.org/W6600424091","https://openalex.org/W6602789111","https://openalex.org/W6602909362","https://openalex.org/W6603599562","https://openalex.org/W6795288823"],"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/W2333966947","https://openalex.org/W2399947890"],"abstract_inverted_index":{"The":[0],"primary":[1],"challenge":[2],"in":[3,7,35,66,209],"text-video":[4],"retrieval":[5,190],"lies":[6],"achieving":[8],"cross-modal":[9,181],"semantic":[10,44,219],"alignment,":[11],"particularly":[12],"the":[13,16,27,33,63,98,106,138,141,149,157],"discrepancy":[14],"between":[15],"conciseness":[17],"of":[18,29,152],"textual":[19,215],"descriptions,":[20,216],"which":[21,38],"often":[22],"fail":[23],"to":[24,80,117,130,155,202],"fully":[25],"encapsulate":[26],"breadth":[28],"video":[30,36,51,123,128,177],"content,":[31],"and":[32,41,50,83,133,176,213,221],"redundancy":[34,132],"data,":[37],"introduces":[39],"noise":[40],"masks":[42],"important":[43],"features.":[45],"Current":[46],"methods":[47,208],"align":[48],"text":[49,115,175],"by":[52,160],"mapping":[53],"them":[54],"into":[55],"a":[56,70,94,102,164],"shared":[57],"feature":[58],"space.":[59],"Despite":[60],"notable":[61],"advancements,":[62],"inherent":[64],"differences":[65],"modality-specific":[67],"representations":[68],"create":[69],"bottleneck":[71],"for":[72],"fixed-point":[73],"embedding":[74,99,158],"techniques,":[75],"making":[76],"models":[77,154],"highly":[78],"sensitive":[79],"dataset":[81],"distribution":[82,167],"hindering":[84],"their":[85],"generalization":[86,222],"ability.":[87],"In":[88,105,137],"this":[89],"paper,":[90],"we":[91,112,126,147],"present":[92],"DomainDiff,":[93],"framework":[95],"that":[96,186],"enhances":[97],"space":[100,159],"through":[101],"two-stage":[103],"process.":[104],"first":[107],"stage,":[108,140],"stochastic":[109],"domain":[110],"modeling,":[111],"semantically":[113],"expand":[114],"embeddings":[116],"explore":[118],"potential":[119],"regions":[120],"aligned":[121],"with":[122,196],"content.":[124],"Simultaneously,":[125],"filter":[127],"segments":[129],"reduce":[131],"highlight":[134],"key":[135],"frames.":[136],"second":[139],"dynamic":[142],"agent":[143,170],"attention":[144,171],"diffusion":[145,153],"network,":[146],"leverage":[148],"generative":[150],"properties":[151],"optimize":[156],"viewing":[161],"it":[162],"from":[163,200],"joint":[165],"probability":[166],"perspective.":[168],"An":[169],"mechanism":[172],"dynamically":[173],"integrates":[174],"features,":[178],"ensuring":[179],"accurate":[180],"alignment.":[182],"Experimental":[183],"results":[184],"demonstrate":[185],"DomainDiff":[187,205],"significantly":[188],"improves":[189],"performance":[191],"across":[192,223],"five":[193],"benchmark":[194],"datasets,":[195],"R@1":[197],"improvements":[198],"ranging":[199],"3%":[201],"7.4%.":[203],"Moreover,":[204],"outperforms":[206],"existing":[207],"handling":[210],"long":[211],"videos":[212],"complex":[214],"showcasing":[217],"superior":[218],"robustness":[220],"varying":[224],"distributions.":[225]},"counts_by_year":[],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-10T00:00:00"}