{"id":"https://openalex.org/W7166660537","doi":"https://doi.org/10.48550/arxiv.2606.30168","title":"Latent Noise Mask for Reducing Visual Redundancy in Multimodal Large Language Models","display_name":"Latent Noise Mask for Reducing Visual Redundancy in Multimodal Large Language Models","publication_year":2026,"publication_date":"2026-06-29","ids":{"openalex":"https://openalex.org/W7166660537","doi":"https://doi.org/10.48550/arxiv.2606.30168"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2606.30168","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.30168","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","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.2606.30168","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5139647180","display_name":"Kai Jiang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jiang, Kai","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5105354608","display_name":"Ry Zhu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhu, Ruishu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5043900893","display_name":"Sicong Huang","orcid":"https://orcid.org/0000-0003-0854-4734"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Huang, Siqi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139670353","display_name":"Hongyuan Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Hongyuan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5139672842","display_name":"Xuelong Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Xuelong","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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.9789000153541565,"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":0.9789000153541565,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.0031999999191612005,"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/T10028","display_name":"Topic Modeling","score":0.0027000000700354576,"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/security-token","display_name":"Security token","score":0.708299994468689},{"id":"https://openalex.org/keywords/redundancy","display_name":"Redundancy (engineering)","score":0.5853000283241272},{"id":"https://openalex.org/keywords/visual-reasoning","display_name":"Visual reasoning","score":0.5325999855995178},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.5001999735832214},{"id":"https://openalex.org/keywords/noise","display_name":"Noise (video)","score":0.41119998693466187},{"id":"https://openalex.org/keywords/overhead","display_name":"Overhead (engineering)","score":0.34929999709129333}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7437999844551086},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.708299994468689},{"id":"https://openalex.org/C152124472","wikidata":"https://www.wikidata.org/wiki/Q1204361","display_name":"Redundancy (engineering)","level":2,"score":0.5853000283241272},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5504999756813049},{"id":"https://openalex.org/C2777508537","wikidata":"https://www.wikidata.org/wiki/Q7936620","display_name":"Visual reasoning","level":2,"score":0.5325999855995178},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.5001999735832214},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.41119998693466187},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.3684000074863434},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.34929999709129333},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.34380000829696655},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.3346000015735626},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.28630000352859497},{"id":"https://openalex.org/C160086991","wikidata":"https://www.wikidata.org/wiki/Q5939193","display_name":"Human visual system model","level":3,"score":0.271699994802475},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.2705000042915344},{"id":"https://openalex.org/C178253425","wikidata":"https://www.wikidata.org/wiki/Q162668","display_name":"Visual perception","level":3,"score":0.259799987077713},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.25940001010894775}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2606.30168","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.30168","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2606.30168","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.30168","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","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":{"Multimodal":[0],"large":[1],"language":[2,31],"models":[3,32],"(MLLMs)":[4],"often":[5],"fail":[6],"in":[7,74],"fine-grained":[8],"visual":[9,13,34,51,61,72,88,171],"reasoning,":[10],"as":[11],"question-relevant":[12,170],"cues":[14,73],"are":[15],"diluted":[16],"by":[17,100,144,152],"dense":[18],"and":[19,94,130,151],"redundant":[20,50],"image":[21],"tokens.":[22,52],"Recent":[23],"multimodal":[24,162],"reasoning":[25,42,163,178],"methods":[26],"usually":[27],"extend":[28],"chain-of-thought":[29],"from":[30,168,174],"into":[33,109],"or":[35,120],"latent":[36,75,107],"spaces,":[37],"seeking":[38],"to":[39,68,85],"add":[40],"intermediate":[41],"states":[43],"while":[44,139],"overlooking":[45],"the":[46,91,101,117,141,177],"negative":[47],"impact":[48],"of":[49],"We":[53],"propose":[54],"LatEnt":[55],"Noise":[56],"maSk":[57],"(Lens),":[58],"a":[59,79,131],"question-conditioned":[60],"evidence":[62,172],"purification":[63],"framework":[64],"that":[65,161],"empowers":[66],"MLLMs":[67],"reason":[69],"with":[70],"cleaner":[71,169],"space.":[76],"Lens":[77,81,135],"introduces":[78],"lightweight":[80,132],"Evidence":[82],"Token":[83],"(LET)":[84],"score":[86],"which":[87],"tokens":[89],"support":[90],"current":[92],"question":[93],"preserve":[95],"them":[96],"during":[97],"decoding.":[98],"Guided":[99],"LET":[102],"scores,":[103],"it":[104],"injects":[105],"adaptive":[106],"noise":[108,133],"low-relevance":[110],"tokens,":[111],"softly":[112],"suppressing":[113],"distractors":[114],"without":[115],"changing":[116],"model":[118],"backbone":[119],"token":[121,129],"sequence.":[122],"With":[123],"only":[124],"one":[125],"temporary":[126],"learnable":[127],"control":[128],"generator,":[134],"adds":[136],"minimal":[137],"overhead":[138],"improving":[140],"base":[142],"MLLM":[143],"2.4-6.4":[145],"points":[146,154],"on":[147,155],"most":[148],"VQA":[149],"datasets":[150],"4.1-6.4":[153],"grounding":[156],"tasks.":[157],"These":[158],"results":[159],"show":[160],"can":[164],"benefit":[165],"more":[166],"directly":[167],"than":[173],"simply":[175],"extending":[176],"trace.":[179]},"counts_by_year":[],"updated_date":"2026-07-01T06:29:00.853634","created_date":"2026-07-01T00:00:00"}