{"id":"https://openalex.org/W4412394955","doi":"https://doi.org/10.1145/3726302.3731945","title":"Embedding-based Retrieval in Multi-Modal Content Moderation","display_name":"Embedding-based Retrieval in Multi-Modal Content Moderation","publication_year":2025,"publication_date":"2025-07-13","ids":{"openalex":"https://openalex.org/W4412394955","doi":"https://doi.org/10.1145/3726302.3731945"},"language":"en","primary_location":{"id":"doi:10.1145/3726302.3731945","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3726302.3731945","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3726302.3731945","source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 48th International ACM SIGIR Conference on Research and Development in Information Retrieval","raw_type":"proceedings-article"},"type":"article","indexed_in":["arxiv","crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://dl.acm.org/doi/pdf/10.1145/3726302.3731945","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":null,"display_name":"Hanzhong Liang","orcid":"https://orcid.org/0009-0007-7101-5292"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hanzhong Liang","raw_affiliation_strings":["TikTok, San Jose, CA, USA"],"raw_orcid":"https://orcid.org/0009-0007-7101-5292","affiliations":[{"raw_affiliation_string":"TikTok, San Jose, CA, USA","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5107901514","display_name":"Jinghao Shi","orcid":"https://orcid.org/0009-0003-0196-5111"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jinghao Shi","raw_affiliation_strings":["TikTok, San Jose, CA, USA"],"raw_orcid":"https://orcid.org/0009-0003-0196-5111","affiliations":[{"raw_affiliation_string":"TikTok, San Jose, CA, USA","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5006504972","display_name":"Xiang Shen","orcid":"https://orcid.org/0000-0002-6457-0411"},"institutions":[{"id":"https://openalex.org/I174135032","display_name":"Bellevue College","ror":"https://ror.org/05gr4yv49","country_code":"US","type":"education","lineage":["https://openalex.org/I174135032"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Xiang Shen","raw_affiliation_strings":["TikTok, Bellevue, WA, USA"],"raw_orcid":"https://orcid.org/0000-0002-6457-0411","affiliations":[{"raw_affiliation_string":"TikTok, Bellevue, WA, USA","institution_ids":["https://openalex.org/I174135032"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Zixuan Wang","orcid":"https://orcid.org/0009-0001-5532-5932"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zixuan Wang","raw_affiliation_strings":["TikTok, San Jose, CA, USA"],"raw_orcid":"https://orcid.org/0009-0001-5532-5932","affiliations":[{"raw_affiliation_string":"TikTok, San Jose, CA, USA","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5109021998","display_name":"Vera Wen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Vera Wen","raw_affiliation_strings":["TikTok, San Jose, CA, USA"],"raw_orcid":"https://orcid.org/0009-0000-7328-409X","affiliations":[{"raw_affiliation_string":"TikTok, San Jose, CA, USA","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5118976561","display_name":"Ardalan Mehrani","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ardalan Mehrani","raw_affiliation_strings":["TikTok, San Jose, CA, USA"],"raw_orcid":"https://orcid.org/0009-0002-1369-7639","affiliations":[{"raw_affiliation_string":"TikTok, San Jose, CA, USA","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5090648171","display_name":"Zhiqian Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhiqian Chen","raw_affiliation_strings":["TikTok, San Jose, CA, USA"],"raw_orcid":"https://orcid.org/0009-0008-7445-4108","affiliations":[{"raw_affiliation_string":"TikTok, San Jose, CA, USA","institution_ids":[]}]},{"author_position":"middle","author":{"id":null,"display_name":"Yifan Wu","orcid":"https://orcid.org/0009-0008-2798-8860"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yifan Wu","raw_affiliation_strings":["TikTok, San Jose, CA, USA"],"raw_orcid":"https://orcid.org/0009-0008-2798-8860","affiliations":[{"raw_affiliation_string":"TikTok, San Jose, CA, USA","institution_ids":[]}]},{"author_position":"last","author":{"id":null,"display_name":"Zhixin Zhang","orcid":"https://orcid.org/0009-0004-3768-7325"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhixin Zhang","raw_affiliation_strings":["TikTok, Singapore, Singapore"],"raw_orcid":"https://orcid.org/0009-0004-3768-7325","affiliations":[{"raw_affiliation_string":"TikTok, Singapore, Singapore","institution_ids":[]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.08073634,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"4264","last_page":"4268"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.9988999962806702,"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"}},"topics":[{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.9988999962806702,"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/T12262","display_name":"Hate Speech and Cyberbullying Detection","score":0.9987999796867371,"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/T10664","display_name":"Sentiment Analysis and Opinion Mining","score":0.9977999925613403,"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/modal","display_name":"Modal","score":0.7646421194076538},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.7607999444007874},{"id":"https://openalex.org/keywords/embedding","display_name":"Embedding","score":0.6761205792427063},{"id":"https://openalex.org/keywords/moderation","display_name":"Moderation","score":0.5842670202255249},{"id":"https://openalex.org/keywords/content","display_name":"Content (measure theory)","score":0.5453932881355286},{"id":"https://openalex.org/keywords/information-retrieval","display_name":"Information retrieval","score":0.4745294153690338},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.3654210865497589},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.14749231934547424},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.11390021443367004}],"concepts":[{"id":"https://openalex.org/C71139939","wikidata":"https://www.wikidata.org/wiki/Q910194","display_name":"Modal","level":2,"score":0.7646421194076538},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7607999444007874},{"id":"https://openalex.org/C41608201","wikidata":"https://www.wikidata.org/wiki/Q980509","display_name":"Embedding","level":2,"score":0.6761205792427063},{"id":"https://openalex.org/C93225998","wikidata":"https://www.wikidata.org/wiki/Q1941972","display_name":"Moderation","level":2,"score":0.5842670202255249},{"id":"https://openalex.org/C2778152352","wikidata":"https://www.wikidata.org/wiki/Q5165061","display_name":"Content (measure theory)","level":2,"score":0.5453932881355286},{"id":"https://openalex.org/C23123220","wikidata":"https://www.wikidata.org/wiki/Q816826","display_name":"Information retrieval","level":1,"score":0.4745294153690338},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3654210865497589},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.14749231934547424},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.11390021443367004},{"id":"https://openalex.org/C134306372","wikidata":"https://www.wikidata.org/wiki/Q7754","display_name":"Mathematical analysis","level":1,"score":0.0},{"id":"https://openalex.org/C188027245","wikidata":"https://www.wikidata.org/wiki/Q750446","display_name":"Polymer chemistry","level":1,"score":0.0},{"id":"https://openalex.org/C185592680","wikidata":"https://www.wikidata.org/wiki/Q2329","display_name":"Chemistry","level":0,"score":0.0}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1145/3726302.3731945","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3726302.3731945","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3726302.3731945","source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 48th International ACM SIGIR Conference on Research and Development in Information Retrieval","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2507.01066","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2507.01066","pdf_url":"https://arxiv.org/pdf/2507.01066","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"}],"best_oa_location":{"id":"doi:10.1145/3726302.3731945","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3726302.3731945","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3726302.3731945","source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 48th International ACM SIGIR Conference on Research and Development in Information Retrieval","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4412394955.pdf","grobid_xml":"https://content.openalex.org/works/W4412394955.grobid-xml"},"referenced_works_count":14,"referenced_works":["https://openalex.org/W2560674852","https://openalex.org/W3203711169","https://openalex.org/W4294653793","https://openalex.org/W4306317484","https://openalex.org/W4380302040","https://openalex.org/W4384648826","https://openalex.org/W4385570624","https://openalex.org/W4386576774","https://openalex.org/W4394593185","https://openalex.org/W4399152175","https://openalex.org/W4403220506","https://openalex.org/W4403322060","https://openalex.org/W4403582492","https://openalex.org/W4409367039"],"related_works":["https://openalex.org/W2740268725","https://openalex.org/W3093710306","https://openalex.org/W750217911","https://openalex.org/W4206949499","https://openalex.org/W2296969234","https://openalex.org/W4226254121","https://openalex.org/W4388229525","https://openalex.org/W131533274","https://openalex.org/W4313163826","https://openalex.org/W1517021018"],"abstract_inverted_index":{"Video":[0],"understanding":[1],"plays":[2],"a":[3,65,73],"fundamental":[4],"role":[5],"for":[6,25],"content":[7,26],"moderation":[8],"on":[9,101,130],"short":[10],"video":[11,118],"platforms,":[12],"enabling":[13],"the":[14,22,109],"detection":[15],"of":[16,75],"inappropriate":[17],"content.":[18],"While":[19],"classification":[20,60],"remains":[21],"dominant":[23],"approach":[24],"moderation,":[27],"it":[28],"often":[29],"struggles":[30],"in":[31],"scenarios":[32],"requiring":[33],"rapid":[34],"and":[35,42,82,98,107,117,123,144,161,172],"cost-efficient":[36],"responses,":[37],"such":[38,95],"as":[39,96],"trend":[40,125],"adaptation":[41],"urgent":[43],"escalations.":[44],"To":[45],"address":[46],"this":[47],"issue,":[48],"we":[49,105],"introduce":[50],"an":[51],"Embedding-Based":[52],"Retrieval":[53],"(EBR)":[54],"method":[55],"designed":[56],"to":[57,71,120,142,148,175],"complement":[58],"traditional":[59],"approaches.":[61],"We":[62],"first":[63],"leverage":[64],"Supervised":[66],"Contrastive":[67],"Learning":[68],"(SCL)":[69],"framework":[70],"train":[72],"suite":[74],"foundation":[76],"embedding":[77,103,115],"models,":[78,104],"including":[79],"both":[80],"single-modal":[81],"multi-modal":[83],"architectures.":[84],"Our":[85],"models":[86],"demonstrate":[87],"superior":[88],"performance":[89],"over":[90,166],"established":[91],"contrastive":[92],"learning":[93],"methods":[94],"CLIP":[97],"MoCo.":[99],"Building":[100],"these":[102],"design":[106],"implement":[108],"embedding-based":[110],"retrieval":[111,119],"system":[112],"that":[113,136,154],"integrates":[114],"generation":[116],"enable":[121],"efficient":[122],"effective":[124],"handling.":[126],"Comprehensive":[127],"offline":[128],"experiments":[129,152],"25":[131],"diverse":[132],"emerging":[133],"trends":[134],"show":[135],"EBR":[137,155],"improves":[138],"ROC-AUC":[139],"from":[140,146],"0.85":[141],"0.99":[143],"PR-AUC":[145],"0.35":[147],"0.95.":[149],"Further":[150],"online":[151],"reveal":[153],"increases":[156],"action":[157],"rates":[158],"by":[159,165],"10.32%":[160],"reduces":[162],"operational":[163],"costs":[164],"80%,":[167],"while":[168],"also":[169],"enhancing":[170],"interpretability":[171],"flexibility":[173],"compared":[174],"classification-based":[176],"solutions.":[177]},"counts_by_year":[],"updated_date":"2026-03-27T05:58:40.876381","created_date":"2025-10-10T00:00:00"}