{"id":"https://openalex.org/W7160641479","doi":"https://doi.org/10.48550/arxiv.2605.05938","title":"ICU-Bench:Benchmarking Continual Unlearning in Multimodal Large Language Models","display_name":"ICU-Bench:Benchmarking Continual Unlearning in Multimodal Large Language Models","publication_year":2026,"publication_date":"2026-05-07","ids":{"openalex":"https://openalex.org/W7160641479","doi":"https://doi.org/10.48550/arxiv.2605.05938"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2605.05938","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.05938","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.2605.05938","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5135688998","display_name":"Yuhang Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Yuhang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135726699","display_name":"Wenjie Mei","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Mei, Wenjie","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135678921","display_name":"Junkai Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Junkai","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135651568","display_name":"Guangyu He","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"He, Guangyu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135666809","display_name":"Zhenxing Niu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Niu, Zhenxing","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5086029723","display_name":"Haichang Gao","orcid":"https://orcid.org/0000-0002-4969-5718"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Gao, Haichang","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/T10028","display_name":"Topic Modeling","score":0.22859999537467957,"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":0.22859999537467957,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.11559999734163284,"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/T10764","display_name":"Privacy-Preserving Technologies in Data","score":0.10199999809265137,"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/forgetting","display_name":"Forgetting","score":0.8184000253677368},{"id":"https://openalex.org/keywords/task","display_name":"Task (project management)","score":0.6100000143051147},{"id":"https://openalex.org/keywords/benchmark","display_name":"Benchmark (surveying)","score":0.4747999906539917},{"id":"https://openalex.org/keywords/bridge","display_name":"Bridge (graph theory)","score":0.46939998865127563},{"id":"https://openalex.org/keywords/scalability","display_name":"Scalability","score":0.45210000872612},{"id":"https://openalex.org/keywords/focus","display_name":"Focus (optics)","score":0.41200000047683716}],"concepts":[{"id":"https://openalex.org/C7149132","wikidata":"https://www.wikidata.org/wiki/Q1377840","display_name":"Forgetting","level":2,"score":0.8184000253677368},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7491999864578247},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.6100000143051147},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5232999920845032},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.4747999906539917},{"id":"https://openalex.org/C100776233","wikidata":"https://www.wikidata.org/wiki/Q2532492","display_name":"Bridge (graph theory)","level":2,"score":0.46939998865127563},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.45210000872612},{"id":"https://openalex.org/C192209626","wikidata":"https://www.wikidata.org/wiki/Q190909","display_name":"Focus (optics)","level":2,"score":0.41200000047683716},{"id":"https://openalex.org/C107457646","wikidata":"https://www.wikidata.org/wiki/Q207434","display_name":"Human\u2013computer interaction","level":1,"score":0.4032000005245209},{"id":"https://openalex.org/C175154964","wikidata":"https://www.wikidata.org/wiki/Q380077","display_name":"Task analysis","level":3,"score":0.3513000011444092},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.3246999979019165},{"id":"https://openalex.org/C2522767166","wikidata":"https://www.wikidata.org/wiki/Q2374463","display_name":"Data science","level":1,"score":0.29660001397132874},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.29440000653266907},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.2872999906539917},{"id":"https://openalex.org/C112972136","wikidata":"https://www.wikidata.org/wiki/Q7595718","display_name":"Stability (learning theory)","level":2,"score":0.2856999933719635},{"id":"https://openalex.org/C184337299","wikidata":"https://www.wikidata.org/wiki/Q1437428","display_name":"Semantics (computer science)","level":2,"score":0.2696000039577484},{"id":"https://openalex.org/C2983448237","wikidata":"https://www.wikidata.org/wiki/Q1078276","display_name":"Language understanding","level":2,"score":0.2517000138759613}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2605.05938","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.05938","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.2605.05938","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.05938","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":{"Although":[0],"Multimodal":[1],"Large":[2],"Language":[3],"Models":[4],"(MLLMs)":[5],"have":[6],"achieved":[7],"remarkable":[8],"progress":[9],"across":[10],"many":[11],"domains,":[12,76],"their":[13],"training":[14],"on":[15,34,64,125],"large-scale":[16],"multimodal":[17,60,159],"datasets":[18],"raises":[19],"serious":[20],"privacy":[21,45,166],"concerns,":[22],"making":[23],"effective":[24],"machine":[25],"unlearning":[26,61,95,116,123,160],"increasingly":[27],"necessary.":[28],"However,":[29],"existing":[30,130],"benchmarks":[31],"mainly":[32],"focus":[33],"static":[35],"or":[36],"short-sequence":[37],"settings,":[38],"offering":[39],"limited":[40],"support":[41],"for":[42,158,164],"evaluating":[43],"continual":[44,59,94,115,135,165],"deletion":[46],"requests":[47],"in":[48,134,141],"realistic":[49],"deployments.":[50],"To":[51],"bridge":[52],"this":[53],"gap,":[54],"we":[55,127],"introduce":[56],"ICU-Bench,":[57,126],"a":[58,100],"benchmark":[62],"built":[63],"privacy-critical":[65],"document":[66,75],"data.":[67],"ICU-Bench":[68],"contains":[69],"1,000":[70],"privacy-sensitive":[71],"profiles":[72],"from":[73],"two":[74],"medical":[77],"reports":[78],"and":[79,88,111,137,147],"labor":[80],"contracts,":[81],"with":[82,121],"9,500":[83],"images,":[84],"16,000":[85],"question-answer":[86],"pairs,":[87],"100":[89],"forget":[90],"tasks.":[91],"Additionally,":[92],"new":[93],"metrics":[96],"are":[97],"introduced,":[98],"facilitating":[99],"comprehensive":[101],"analysis":[102],"of":[103],"forgetting":[104,107,143],"effectiveness,":[105],"historical":[106],"preservation,":[108,146],"retained":[109],"utility,":[110],"stability":[112],"throughout":[113],"the":[114,156],"process.":[117],"Through":[118],"extensive":[119],"experiments":[120],"representative":[122],"methods":[124,131,161],"show":[128],"that":[129],"generally":[132],"struggle":[133],"settings":[136],"exhibit":[138],"clear":[139],"limitations":[140],"balancing":[142],"quality,":[144],"utility":[145],"scalability":[148],"over":[149],"long":[150],"task":[151],"sequences.":[152],"These":[153],"findings":[154],"highlight":[155],"need":[157],"explicitly":[162],"designed":[163],"deletion.":[167]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-05-09T00:00:00"}