{"id":"https://openalex.org/W7161964087","doi":"https://doi.org/10.48550/arxiv.2605.20584","title":"QwenSafe: Multimodal Content Rating Description Identification via Preference-Aligned VLMs","display_name":"QwenSafe: Multimodal Content Rating Description Identification via Preference-Aligned VLMs","publication_year":2026,"publication_date":"2026-05-20","ids":{"openalex":"https://openalex.org/W7161964087","doi":"https://doi.org/10.48550/arxiv.2605.20584"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2605.20584","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.20584","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":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.2605.20584","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5083901377","display_name":"Dishanika Denipitiyage","orcid":"https://orcid.org/0000-0002-8717-098X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Denipitiyage, Dishanika","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134113377","display_name":"Aruna Seneviratne","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Seneviratne, Aruna","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5038376039","display_name":"Suranga Seneviratne","orcid":"https://orcid.org/0000-0002-5485-5595"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Seneviratne, Suranga","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/T10260","display_name":"Software Engineering Research","score":0.1264999955892563,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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/T10260","display_name":"Software Engineering Research","score":0.1264999955892563,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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/T10028","display_name":"Topic Modeling","score":0.07980000227689743,"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":0.07450000196695328,"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/consistency","display_name":"Consistency (knowledge bases)","score":0.6784999966621399},{"id":"https://openalex.org/keywords/metadata","display_name":"Metadata","score":0.6272000074386597},{"id":"https://openalex.org/keywords/scalability","display_name":"Scalability","score":0.60589998960495},{"id":"https://openalex.org/keywords/precision-and-recall","display_name":"Precision and recall","score":0.5602999925613403},{"id":"https://openalex.org/keywords/pipeline","display_name":"Pipeline (software)","score":0.5133000016212463},{"id":"https://openalex.org/keywords/recall","display_name":"Recall","score":0.42989999055862427},{"id":"https://openalex.org/keywords/identification","display_name":"Identification (biology)","score":0.366100013256073},{"id":"https://openalex.org/keywords/mobile-device","display_name":"Mobile device","score":0.35670000314712524}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7739999890327454},{"id":"https://openalex.org/C2776436953","wikidata":"https://www.wikidata.org/wiki/Q5163215","display_name":"Consistency (knowledge bases)","level":2,"score":0.6784999966621399},{"id":"https://openalex.org/C93518851","wikidata":"https://www.wikidata.org/wiki/Q180160","display_name":"Metadata","level":2,"score":0.6272000074386597},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.60589998960495},{"id":"https://openalex.org/C81669768","wikidata":"https://www.wikidata.org/wiki/Q2359161","display_name":"Precision and recall","level":2,"score":0.5602999925613403},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5600000023841858},{"id":"https://openalex.org/C43521106","wikidata":"https://www.wikidata.org/wiki/Q2165493","display_name":"Pipeline (software)","level":2,"score":0.5133000016212463},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5080999732017517},{"id":"https://openalex.org/C23123220","wikidata":"https://www.wikidata.org/wiki/Q816826","display_name":"Information retrieval","level":1,"score":0.5078999996185303},{"id":"https://openalex.org/C100660578","wikidata":"https://www.wikidata.org/wiki/Q18733","display_name":"Recall","level":2,"score":0.42989999055862427},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.3781999945640564},{"id":"https://openalex.org/C116834253","wikidata":"https://www.wikidata.org/wiki/Q2039217","display_name":"Identification (biology)","level":2,"score":0.366100013256073},{"id":"https://openalex.org/C186967261","wikidata":"https://www.wikidata.org/wiki/Q5082128","display_name":"Mobile device","level":2,"score":0.35670000314712524},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.34200000762939453},{"id":"https://openalex.org/C2778152352","wikidata":"https://www.wikidata.org/wiki/Q5165061","display_name":"Content (measure theory)","level":2,"score":0.32499998807907104},{"id":"https://openalex.org/C36464697","wikidata":"https://www.wikidata.org/wiki/Q451553","display_name":"Visualization","level":2,"score":0.3212999999523163},{"id":"https://openalex.org/C148524875","wikidata":"https://www.wikidata.org/wiki/Q6975395","display_name":"F1 score","level":2,"score":0.31220000982284546},{"id":"https://openalex.org/C2988145974","wikidata":"https://www.wikidata.org/wiki/Q620615","display_name":"Mobile apps","level":2,"score":0.3118000030517578},{"id":"https://openalex.org/C66905080","wikidata":"https://www.wikidata.org/wiki/Q17005494","display_name":"Binary classification","level":3,"score":0.30570000410079956},{"id":"https://openalex.org/C171686336","wikidata":"https://www.wikidata.org/wiki/Q3532085","display_name":"Topic model","level":2,"score":0.30559998750686646},{"id":"https://openalex.org/C2781249084","wikidata":"https://www.wikidata.org/wiki/Q908656","display_name":"Preference","level":2,"score":0.29829999804496765},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.28870001435279846},{"id":"https://openalex.org/C2778598663","wikidata":"https://www.wikidata.org/wiki/Q1407599","display_name":"Video content analysis","level":4,"score":0.2662999927997589}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2605.20584","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.20584","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":"doi:10.48550/arxiv.2605.20584","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.20584","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":false,"raw_source_name":null,"raw_type":"Preprint"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Mobile":[0],"app":[1,37,69,93,195],"marketplaces":[2],"require":[3],"developers":[4],"to":[5,12,32,57,112,186],"disclose":[6],"standardized":[7],"content":[8,132,177,191],"rating":[9,133,192],"descriptors":[10,134],"(CRDs)":[11],"inform":[13],"users":[14],"about":[15],"potentially":[16],"sensitive":[17],"or":[18],"restricted":[19],"content.":[20],"Ensuring":[21],"the":[22,33,60,181],"accuracy":[23],"and":[24,43,71,96,119,123,135,145,164,179,189],"consistency":[25],"of":[26,36,62,161,183],"these":[27],"disclosures":[28],"remains":[29],"challenging":[30],"due":[31],"multimodal":[34,172],"nature":[35],"content,":[38],"which":[39],"spans":[40],"textual":[41,124],"descriptions":[42],"visual":[44,122],"interfaces.":[45],"In":[46],"this":[47,78],"paper,":[48],"we":[49,80],"present":[50],"QwenSafe,":[51],"a":[52,83],"Vision-Language":[53],"Model":[54],"(VLM)":[55],"designed":[56],"automatically":[58],"identify":[59],"presence":[61],"Apple-defined":[63,131],"CRDs":[64],"by":[65,91,107],"jointly":[66],"reasoning":[67],"over":[68],"metadata":[70],"screenshots.":[72],"To":[73],"enable":[74],"scalable":[75,188],"training":[76],"for":[77],"task,":[79],"introduce":[81],"metadata2CRD,":[82],"data-construction":[84],"pipeline":[85],"that":[86,170],"synthesizes":[87],"descriptor-aligned":[88],"question-answer":[89],"pairs":[90],"combining":[92],"descriptions,":[94],"screenshots,":[95],"formal":[97],"descriptor":[98],"definitions.":[99],"We":[100,126],"adapt":[101],"Qwen3-VL-8B":[102],"using":[103],"supervised":[104],"fine-tuning":[105],"followed":[106],"Direct":[108],"Preference":[109],"Optimization":[110],"(DPO)":[111],"align":[113],"model":[114],"predictions":[115],"with":[116],"descriptor-specific":[117],"evidence":[118],"explanations":[120],"across":[121],"modalities.":[125],"evaluate":[127],"QwenSafe":[128,147],"on":[129],"12":[130],"compare":[136],"it":[137],"against":[138],"state-of-the-art":[139],"vision-language":[140,184],"models,":[141],"including":[142],"Qwen3-VL,":[143],"LLaVA-1.6,":[144],"Gemini-2.5-Flash.":[146],"consistently":[148],"outperforms":[149],"all":[150],"baselines":[151],"in":[152,158,193],"binary":[153],"CRD":[154],"classification,":[155],"achieving":[156],"improvements":[157],"positive-class":[159],"recall":[160],"111.8%,":[162],"36.1%,":[163],"2.1%,":[165],"respectively.":[166],"Our":[167],"results":[168],"demonstrate":[169],"descriptor-aware":[171],"alignment":[173],"substantially":[174],"improves":[175],"automated":[176],"classification":[178],"highlights":[180],"potential":[182],"models":[185],"support":[187],"consistent":[190],"mobile":[194],"marketplaces.":[196]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-05-22T00:00:00"}