{"id":"https://openalex.org/W7160936012","doi":"https://doi.org/10.48550/arxiv.2605.10386","title":"GuardAD: Safeguarding Autonomous Driving MLLMs via Markovian Safety Logic","display_name":"GuardAD: Safeguarding Autonomous Driving MLLMs via Markovian Safety Logic","publication_year":2026,"publication_date":"2026-05-11","ids":{"openalex":"https://openalex.org/W7160936012","doi":"https://doi.org/10.48550/arxiv.2605.10386"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2605.10386","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.10386","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.2605.10386","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5135983069","display_name":"Tianyuan Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Tianyuan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5003200822","display_name":"Peng Yue","orcid":"https://orcid.org/0000-0003-3006-4542"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yue, Peng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5036466780","display_name":"Zihao Peng","orcid":"https://orcid.org/0000-0003-3108-123X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Peng, Zihao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5084646389","display_name":"J.H. Liu","orcid":"https://orcid.org/0000-0002-1968-150X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Jiangfan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5020945039","display_name":"Zonghao Ying","orcid":"https://orcid.org/0009-0007-7393-7362"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ying, Zonghao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135927119","display_name":"Jiakai Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Jiakai","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135928206","display_name":"Tianlin Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Tianlin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135981891","display_name":"Jian Yang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yang, Jian","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135922059","display_name":"Yaodong Yang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yang, Yaodong","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135957514","display_name":"Aishan Liu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Aishan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5135948096","display_name":"Xianglong Liu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Xianglong","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/T11099","display_name":"Autonomous Vehicle Technology and Safety","score":0.4909999966621399,"subfield":{"id":"https://openalex.org/subfields/2203","display_name":"Automotive 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/T11099","display_name":"Autonomous Vehicle Technology and Safety","score":0.4909999966621399,"subfield":{"id":"https://openalex.org/subfields/2203","display_name":"Automotive 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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.36079999804496765,"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":0.013299999758601189,"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/safeguard","display_name":"Safeguard","score":0.589900016784668},{"id":"https://openalex.org/keywords/robustness","display_name":"Robustness (evolution)","score":0.5023999810218811},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.4991999864578247},{"id":"https://openalex.org/keywords/action","display_name":"Action (physics)","score":0.4959000051021576},{"id":"https://openalex.org/keywords/safeguarding","display_name":"Safeguarding","score":0.45910000801086426},{"id":"https://openalex.org/keywords/markov-process","display_name":"Markov process","score":0.43779999017715454},{"id":"https://openalex.org/keywords/safer","display_name":"SAFER","score":0.4374000132083893},{"id":"https://openalex.org/keywords/task","display_name":"Task (project management)","score":0.43290001153945923}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.64410001039505},{"id":"https://openalex.org/C2780771206","wikidata":"https://www.wikidata.org/wiki/Q3271761","display_name":"Safeguard","level":2,"score":0.589900016784668},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.5023999810218811},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.4991999864578247},{"id":"https://openalex.org/C2780791683","wikidata":"https://www.wikidata.org/wiki/Q846785","display_name":"Action (physics)","level":2,"score":0.4959000051021576},{"id":"https://openalex.org/C2776743756","wikidata":"https://www.wikidata.org/wiki/Q5097921","display_name":"Safeguarding","level":2,"score":0.45910000801086426},{"id":"https://openalex.org/C112930515","wikidata":"https://www.wikidata.org/wiki/Q4389547","display_name":"Risk analysis (engineering)","level":1,"score":0.4544000029563904},{"id":"https://openalex.org/C159886148","wikidata":"https://www.wikidata.org/wiki/Q176645","display_name":"Markov process","level":2,"score":0.43779999017715454},{"id":"https://openalex.org/C2776654903","wikidata":"https://www.wikidata.org/wiki/Q2601463","display_name":"SAFER","level":2,"score":0.4374000132083893},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.43290001153945923},{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","level":1,"score":0.38280001282691956},{"id":"https://openalex.org/C87833898","wikidata":"https://www.wikidata.org/wiki/Q1060280","display_name":"Advanced driver assistance systems","level":2,"score":0.34450000524520874},{"id":"https://openalex.org/C132835097","wikidata":"https://www.wikidata.org/wiki/Q7663745","display_name":"System safety","level":2,"score":0.3280999958515167},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.32089999318122864},{"id":"https://openalex.org/C127757376","wikidata":"https://www.wikidata.org/wiki/Q2056514","display_name":"Active safety","level":2,"score":0.3158000111579895},{"id":"https://openalex.org/C182306322","wikidata":"https://www.wikidata.org/wiki/Q1779371","display_name":"Order (exchange)","level":2,"score":0.2921999990940094},{"id":"https://openalex.org/C2776544517","wikidata":"https://www.wikidata.org/wiki/Q189447","display_name":"Unexpected events","level":2,"score":0.27810001373291016},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.27140000462532043},{"id":"https://openalex.org/C21847791","wikidata":"https://www.wikidata.org/wiki/Q191081","display_name":"Logical conjunction","level":2,"score":0.2605000138282776},{"id":"https://openalex.org/C3017944768","wikidata":"https://www.wikidata.org/wiki/Q1450463","display_name":"Poison control","level":2,"score":0.26019999384880066}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2605.10386","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.10386","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.2605.10386","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.10386","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":[{"id":"https://metadata.un.org/sdg/3","score":0.690660834312439,"display_name":"Good health and well-being"}],"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],"models":[3],"(MLLMs)":[4],"are":[5],"increasingly":[6],"integrated":[7],"into":[8],"autonomous":[9],"driving":[10,57],"(AD)":[11],"systems;":[12],"however,":[13],"they":[14],"remain":[15],"vulnerable":[16],"to":[17],"diverse":[18],"safety":[19,45,72,86,128],"threats,":[20],"particularly":[21],"in":[22,52,55],"accident-prone":[23],"scenarios.":[24],"Recent":[25],"safeguard":[26,68],"mechanisms":[27],"have":[28],"shown":[29],"promise":[30],"by":[31],"incorporating":[32],"logical":[33,77],"constraints,":[34],"yet":[35],"most":[36],"rely":[37],"on":[38,141],"static":[39],"formulations":[40],"that":[41,69,147],"lack":[42],"temporally":[43],"grounded":[44],"reasoning":[46],"over":[47,88],"evolving":[48,75],"traffic":[49,90],"interactions,":[50],"resulting":[51],"limited":[53],"robustness":[54],"dynamic":[56],"environments.":[58],"To":[59],"address":[60],"these":[61],"limitations,":[62],"we":[63],"propose":[64],"GuardAD,":[65],"a":[66],"model-agnostic":[67],"formulates":[70],"AD":[71],"as":[73],"an":[74],"Markovian":[76,99],"state.":[78],"GuardAD":[79,121,148],"introduces":[80],"Neuro-Symbolic":[81],"Logic":[82,100],"Formalization,":[83],"which":[84],"represents":[85],"predicates":[87],"heterogeneous":[89],"participants":[91],"and":[92,109,144,173],"continuously":[93],"induces":[94],"them":[95],"via":[96],"n-th":[97],"order":[98],"Induction.":[101],"This":[102],"design":[103],"enables":[104],"the":[105,136,171],"inference":[106],"of":[107,175],"emerging":[108],"latent":[110],"hazards":[111],"beyond":[112],"single-step":[113],"observations.":[114],"Rather":[115],"than":[116],"simply":[117],"vetoing":[118],"unsafe":[119],"actions,":[120],"performs":[122],"Logic-Driven":[123],"Action":[124],"Revision,":[125],"where":[126],"inferred":[127],"states":[129],"actively":[130],"guide":[131],"action":[132],"refinement":[133],"without":[134],"modifying":[135],"underlying":[137],"MLLM.":[138],"Extensive":[139],"experiments":[140],"multiple":[142],"benchmarks":[143],"AD-MLLMs":[145],"demonstrate":[146],"substantially":[149],"reduces":[150],"accident":[151],"rates":[152],"(-32.07%)":[153],"while":[154],"slightly":[155],"improving":[156],"task":[157],"performance":[158],"(+6.85%).":[159],"Moreover,":[160],"closed-loop":[161],"simulation":[162],"evaluations,":[163],"together":[164],"with":[165],"physical-world":[166],"vehicle":[167],"studies,":[168],"further":[169],"validate":[170],"effectiveness":[172],"potential":[174],"GuardAD.":[176]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-05-13T00:00:00"}