{"id":"https://openalex.org/W7127440293","doi":"https://doi.org/10.48550/arxiv.2602.01158","title":"Improving Robustness of Vision-Language-Action Models by Restoring Corrupted Visual Inputs","display_name":"Improving Robustness of Vision-Language-Action Models by Restoring Corrupted Visual Inputs","publication_year":2026,"publication_date":"2026-02-01","ids":{"openalex":"https://openalex.org/W7127440293","doi":"https://doi.org/10.48550/arxiv.2602.01158"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2602.01158","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2602.01158","pdf_url":"https://arxiv.org/pdf/2602.01158","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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"},"volume":null,"type":"preprint","indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2602.01158","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5124877125","display_name":"Daniel Yezid Guarnizo Orjuela","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Orjuela, Daniel Yezid Guarnizo","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5124971122","display_name":"Leonardo Scappatura","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Scappatura, Leonardo","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5124876186","display_name":"Veronica Di Gennaro","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Di Gennaro, Veronica","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5124882136","display_name":"Riccardo Andrea Izzo","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Izzo, Riccardo Andrea","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5017486356","display_name":"Gianluca Bardaro","orcid":"https://orcid.org/0000-0002-6695-0012","observed_orcids":["https://orcid.org/0000-0002-6695-0012"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Bardaro, Gianluca","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5003932703","display_name":"Matteo Matteucci","orcid":"https://orcid.org/0000-0002-8306-6739","observed_orcids":["https://orcid.org/0000-0002-8306-6739","https://orcid.org/0009-0001-3286-3236"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Matteucci, Matteo","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/T10653","display_name":"Robot Manipulation and Learning","score":0.9789999723434448,"subfield":{"id":"https://openalex.org/subfields/2207","display_name":"Control and Systems Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T10653","display_name":"Robot Manipulation and Learning","score":0.9789999723434448,"subfield":{"id":"https://openalex.org/subfields/2207","display_name":"Control and Systems Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10462","display_name":"Reinforcement Learning in Robotics","score":0.008299999870359898,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.003599999938160181,"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/vision-language-action-models","display_name":"vision-language-action models","score":0.9919999837875366},{"id":"https://openalex.org/keywords/image-corruptions","display_name":"image corruptions","score":0.7289999723434448},{"id":"https://openalex.org/keywords/sensor-noise","display_name":"sensor noise","score":0.45100000500679016},{"id":"https://openalex.org/keywords/dead-pixels","display_name":"dead pixels","score":0.5920000076293945},{"id":"https://openalex.org/keywords/lens-contamination","display_name":"lens contamination","score":0.796999990940094},{"id":"https://openalex.org/keywords/robustness","display_name":"robustness","score":0.527999997138977},{"id":"https://openalex.org/keywords/vision-transformer","display_name":"vision transformer","score":0.7839999794960022},{"id":"https://openalex.org/keywords/adversarial-training","display_name":"adversarial training","score":0.9800000190734863},{"id":"https://openalex.org/keywords/robotic-manipulation","display_name":"robotic manipulation","score":0.6949999928474426},{"id":"https://openalex.org/keywords/libero","display_name":"LIBERO","score":0.5509999990463257},{"id":"https://openalex.org/keywords/meta-world","display_name":"Meta-World","score":0.9700000286102295}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7074999809265137},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.6776000261306763},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5776000022888184},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.538100004196167},{"id":"https://openalex.org/C105339364","wikidata":"https://www.wikidata.org/wiki/Q2297740","display_name":"Software deployment","level":2,"score":0.49160000681877136},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.43549999594688416},{"id":"https://openalex.org/C160633673","wikidata":"https://www.wikidata.org/wiki/Q355198","display_name":"Pixel","level":2,"score":0.39579999446868896},{"id":"https://openalex.org/C37736160","wikidata":"https://www.wikidata.org/wiki/Q1801315","display_name":"Adversarial system","level":2,"score":0.383899986743927},{"id":"https://openalex.org/C2780154230","wikidata":"https://www.wikidata.org/wiki/Q513420","display_name":"Undo","level":2,"score":0.37599998712539673},{"id":"https://openalex.org/C2780909371","wikidata":"https://www.wikidata.org/wiki/Q4801092","display_name":"Artificial noise","level":4,"score":0.3601999878883362},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.3465999960899353},{"id":"https://openalex.org/C115051666","wikidata":"https://www.wikidata.org/wiki/Q6522493","display_name":"Ranging","level":2,"score":0.34139999747276306},{"id":"https://openalex.org/C26760741","wikidata":"https://www.wikidata.org/wiki/Q160402","display_name":"Perception","level":2,"score":0.3361000120639801},{"id":"https://openalex.org/C106430172","wikidata":"https://www.wikidata.org/wiki/Q6002272","display_name":"Image restoration","level":4,"score":0.33169999718666077},{"id":"https://openalex.org/C113364801","wikidata":"https://www.wikidata.org/wiki/Q26674","display_name":"High fidelity","level":2,"score":0.3240000009536743},{"id":"https://openalex.org/C36464697","wikidata":"https://www.wikidata.org/wiki/Q451553","display_name":"Visualization","level":2,"score":0.31610000133514404},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.3127000033855438},{"id":"https://openalex.org/C140547941","wikidata":"https://www.wikidata.org/wiki/Q7797194","display_name":"Threat model","level":2,"score":0.3086000084877014},{"id":"https://openalex.org/C80191262","wikidata":"https://www.wikidata.org/wiki/Q5477668","display_name":"Fragility","level":2,"score":0.30320000648498535},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.28040000796318054},{"id":"https://openalex.org/C178253425","wikidata":"https://www.wikidata.org/wiki/Q162668","display_name":"Visual perception","level":3,"score":0.2797999978065491},{"id":"https://openalex.org/C112987892","wikidata":"https://www.wikidata.org/wiki/Q5051574","display_name":"Catastrophic failure","level":2,"score":0.27410000562667847},{"id":"https://openalex.org/C2776459999","wikidata":"https://www.wikidata.org/wiki/Q2119376","display_name":"Fidelity","level":2,"score":0.25940001010894775}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2602.01158","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2602.01158","pdf_url":"https://arxiv.org/pdf/2602.01158","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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"},{"id":"doi:10.48550/arxiv.2602.01158","doi":"https://doi.org/10.48550/arxiv.2602.01158","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2602.01158","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2602.01158","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2602.01158","pdf_url":"https://arxiv.org/pdf/2602.01158","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W7223140919","https://openalex.org/W7161707951","https://openalex.org/W7084389655","https://openalex.org/W4415276858","https://openalex.org/W6966695030","https://openalex.org/W4415275278","https://openalex.org/W4413925127","https://openalex.org/W7154313386","https://openalex.org/W7214724283","https://openalex.org/W7164186804"],"abstract_inverted_index":{"Vision-Language-Action":[0],"(VLA)":[1],"models":[2,138],"have":[3],"emerged":[4],"as":[5,96,110,112],"a":[6,18,52,128],"dominant":[7],"paradigm":[8],"for":[9],"generalist":[10],"robotic":[11],"manipulation,":[12],"unifying":[13],"perception":[14],"and":[15,67,98,130,168],"control":[16],"within":[17],"single":[19],"end-to-end":[20],"architecture.":[21],"However,":[22],"despite":[23],"their":[24,36],"success":[25,107,183],"in":[26],"controlled":[27],"environments,":[28],"reliable":[29],"real-world":[30],"deployment":[31],"is":[32],"severely":[33],"hindered":[34],"by":[35,49],"fragility":[37],"to":[38,70,82,109,135,180],"visual":[39,79,188],"disturbances.":[40,141],"While":[41],"existing":[42],"literature":[43],"extensively":[44],"addresses":[45],"physical":[46],"occlusions":[47],"caused":[48],"scene":[50],"geometry,":[51],"critical":[53],"mode":[54],"remains":[55],"largely":[56],"unexplored:":[57],"image":[58],"corruptions.":[59],"These":[60],"sensor-level":[61],"artifacts,":[62],"ranging":[63],"from":[64,105,151],"electronic":[65],"noise":[66],"dead":[68],"pixels":[69],"lens":[71],"contaminants,":[72],"directly":[73],"compromise":[74],"the":[75,78,123,160,166],"integrity":[76],"of":[77,159],"signal":[80,116],"prior":[81],"interpretation.":[83],"In":[84],"this":[85,89],"work,":[86],"we":[87,121],"quantify":[88],"vulnerability,":[90],"demonstrating":[91],"that":[92,172],"state-of-the-art":[93],"VLAs":[94,179],"such":[95],"$\u03c0_{0.5}$":[97],"SmolVLA,":[99],"suffer":[100],"catastrophic":[101],"performance":[102],"degradation,":[103],"dropping":[104],"90\\%":[106],"rates":[108],"low":[111],"2\\%,":[113],"under":[114,186],"common":[115],"artifacts.":[117],"To":[118],"mitigate":[119],"this,":[120],"introduce":[122],"Corruption":[124],"Restoration":[125],"Transformer":[126],"(CRT),":[127],"plug-and-play":[129],"model-agnostic":[131],"vision":[132],"transformer":[133],"designed":[134],"immunize":[136],"VLA":[137],"against":[139],"sensor":[140],"Leveraging":[142],"an":[143],"adversarial":[144],"training":[145],"objective,":[146],"CRT":[147,173],"restores":[148],"clean":[149],"observations":[150],"corrupted":[152],"inputs":[153],"without":[154],"requiring":[155],"computationally":[156],"expensive":[157],"fine-tuning":[158],"underlying":[161],"model.":[162],"Extensive":[163],"experiments":[164],"across":[165],"LIBERO":[167],"Meta-World":[169],"benchmarks":[170],"demonstrate":[171],"effectively":[174],"recovers":[175],"lost":[176],"performance,":[177],"enabling":[178],"maintain":[181],"near-baseline":[182],"rates,":[184],"even":[185],"severe":[187],"corruption.":[189]},"counts_by_year":[],"updated_date":"2026-10-28T10:33:24.094632","created_date":"2026-02-04T00:00:00"}