{"id":"https://openalex.org/W7216201735","doi":"https://doi.org/10.48550/arxiv.2610.00706","title":"AnchorPrompt: Self-Distilled Soft Prompts for Robust Audio-Language Models","display_name":"AnchorPrompt: Self-Distilled Soft Prompts for Robust Audio-Language Models","publication_year":2026,"publication_date":"2026-09-30","ids":{"openalex":"https://openalex.org/W7216201735","doi":"https://doi.org/10.48550/arxiv.2610.00706"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2610.00706","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.00706","pdf_url":"https://arxiv.org/pdf/2610.00706","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","is_traditional":true,"indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2610.00706","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5038687274","display_name":"Pooneh Mousavi","orcid":"https://orcid.org/0009-0007-5761-5988","observed_orcids":["https://orcid.org/0009-0007-5761-5988"]},"institutions":[{"id":"https://openalex.org/I4210164802","display_name":"Mila - Quebec Artificial Intelligence Institute","ror":"https://ror.org/05c22rx21","country_code":"CA","type":"facility","lineage":["https://openalex.org/I4210164802"]},{"id":"https://openalex.org/I60158472","display_name":"Concordia University","ror":"https://ror.org/0420zvk78","country_code":"CA","type":"education","lineage":["https://openalex.org/I60158472"]}],"countries":["CA"],"is_corresponding":false,"raw_author_name":"Mousavi, Pooneh","raw_affiliation_strings":["Concordia University","Mila -Quebec AI Institute"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Concordia University","institution_ids":["https://openalex.org/I60158472"]},{"raw_affiliation_string":"Mila -Quebec AI Institute","institution_ids":["https://openalex.org/I4210164802"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5080988947","display_name":"Amir Ivry","orcid":"https://orcid.org/0000-0002-9043-7040","observed_orcids":["https://orcid.org/0000-0002-9043-7040"]},"institutions":[{"id":"https://openalex.org/I174306211","display_name":"Technion \u2013 Israel Institute of Technology","ror":"https://ror.org/03qryx823","country_code":"IL","type":"education","lineage":["https://openalex.org/I174306211"]}],"countries":["IL"],"is_corresponding":false,"raw_author_name":"Ivry, Amir","raw_affiliation_strings":["Technion -IIT"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Technion -IIT","institution_ids":["https://openalex.org/I174306211"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5040811098","display_name":"Mirco Ravanelli","orcid":"https://orcid.org/0000-0002-3929-5526","observed_orcids":["https://orcid.org/0000-0002-3929-5526","https://orcid.org/0000-0001-8479-3121"]},"institutions":[{"id":"https://openalex.org/I4210164802","display_name":"Mila - Quebec Artificial Intelligence Institute","ror":"https://ror.org/05c22rx21","country_code":"CA","type":"facility","lineage":["https://openalex.org/I4210164802"]},{"id":"https://openalex.org/I60158472","display_name":"Concordia University","ror":"https://ror.org/0420zvk78","country_code":"CA","type":"education","lineage":["https://openalex.org/I60158472"]}],"countries":["CA"],"is_corresponding":false,"raw_author_name":"Ravanelli, Mirco","raw_affiliation_strings":["Concordia University","Mila -Quebec AI Institute"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Concordia University","institution_ids":["https://openalex.org/I60158472"]},{"raw_affiliation_string":"Mila -Quebec AI Institute","institution_ids":["https://openalex.org/I4210164802"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5023830739","display_name":"Cem Subakan","orcid":"https://orcid.org/0000-0002-7593-6589","observed_orcids":["https://orcid.org/0000-0002-7593-6589"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Subakan, Cem","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":3,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.28130000829696655,"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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.28130000829696655,"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/T10201","display_name":"Speech Recognition and Synthesis","score":0.2687999904155731,"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":0.156700000166893,"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/audio-language-models","display_name":"audio-language models","score":0.7580000162124634},{"id":"https://openalex.org/keywords/soft-prompts","display_name":"soft prompts","score":0.8510000109672546},{"id":"https://openalex.org/keywords/self-distillation","display_name":"self-distillation","score":0.9819999933242798},{"id":"https://openalex.org/keywords/adversarial-perturbations","display_name":"adversarial perturbations","score":0.6380000114440918},{"id":"https://openalex.org/keywords/hallucination-mitigation","display_name":"hallucination mitigation","score":0.7519999742507935},{"id":"https://openalex.org/keywords/answer-consistency","display_name":"answer consistency","score":0.8040000200271606},{"id":"https://openalex.org/keywords/zero-shot-transfer","display_name":"zero-shot transfer","score":0.6620000004768372}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7071999907493591},{"id":"https://openalex.org/C2776436953","wikidata":"https://www.wikidata.org/wiki/Q5163215","display_name":"Consistency (knowledge bases)","level":2,"score":0.6816999912261963},{"id":"https://openalex.org/C37736160","wikidata":"https://www.wikidata.org/wiki/Q1801315","display_name":"Adversarial system","level":2,"score":0.5325999855995178},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.49549999833106995},{"id":"https://openalex.org/C2777210771","wikidata":"https://www.wikidata.org/wiki/Q4927124","display_name":"Block (permutation group theory)","level":2,"score":0.4702000021934509},{"id":"https://openalex.org/C2776175482","wikidata":"https://www.wikidata.org/wiki/Q1195816","display_name":"Transfer (computing)","level":2,"score":0.4406999945640564},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.43630000948905945},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.35089999437332153},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.3499000072479248},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.31540000438690186},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.310699999332428},{"id":"https://openalex.org/C39613435","wikidata":"https://www.wikidata.org/wiki/Q846677","display_name":"Remainder","level":2,"score":0.31040000915527344},{"id":"https://openalex.org/C197424946","wikidata":"https://www.wikidata.org/wiki/Q1165717","display_name":"Waveform","level":3,"score":0.2939000129699707},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.2897000014781952},{"id":"https://openalex.org/C150899416","wikidata":"https://www.wikidata.org/wiki/Q1820378","display_name":"Transfer of learning","level":2,"score":0.2888000011444092},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.26809999346733093}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2610.00706","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.00706","pdf_url":"https://arxiv.org/pdf/2610.00706","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.2610.00706","doi":"https://doi.org/10.48550/arxiv.2610.00706","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2610.00706","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:2610.00706","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.00706","pdf_url":"https://arxiv.org/pdf/2610.00706","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":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7216201735.pdf","grobid_xml":"https://content.openalex.org/works/W7216201735.grobid-xml"},"referenced_works_count":7,"referenced_works":["https://openalex.org/W4400717319","https://openalex.org/W4409051416","https://openalex.org/W4415255153","https://openalex.org/W4416381162","https://openalex.org/W7155201584","https://openalex.org/W7160623959","https://openalex.org/W7162411075"],"related_works":["https://openalex.org/W7162411075","https://openalex.org/W7167099738","https://openalex.org/W7154742540","https://openalex.org/W4415273575","https://openalex.org/W7220663755","https://openalex.org/W7148362664","https://openalex.org/W7118021772","https://openalex.org/W4415433791","https://openalex.org/W4414588183","https://openalex.org/W4406794295"],"abstract_inverted_index":{"Large":[0],"audio-language":[1],"models":[2],"(LALMs)":[3],"are":[4],"sensitive":[5],"to":[6,93,111,172],"input":[7],"perturbations,":[8,174],"such":[9,175],"as":[10,76,176],"noise,":[11],"waveform":[12],"corruption,":[13],"and":[14,29,45,57,64,82,107,121,179],"adversarial":[15],"injections.":[16],"We":[17,48,114],"propose":[18],"AnchorPrompt,":[19],"an":[20],"efficient":[21],"adaptation":[22],"method":[23],"that":[24,123],"keeps":[25],"the":[26,39,43,69,73,77,88,145],"model":[27],"frozen":[28],"learns":[30],"a":[31,84],"single":[32],"block":[33],"of":[34,105,137,147],"prompt":[35],"vectors":[36,51],"inserted":[37],"at":[38,99,149],"decoder":[40],"input,":[41],"between":[42],"audio":[44,56,89,157,165],"question":[46],"embeddings.":[47],"train":[49],"these":[50,168],"through":[52],"self-distillation":[53],"over":[54],"diverse":[55],"text":[58],"perturbations.":[59],"To":[60],"improve":[61],"answer":[62,126],"consistency":[63,127,169],"mitigate":[65],"hallucination,":[66],"we":[67],"use":[68],"model's":[70],"prediction":[71],"on":[72,144,163],"clean":[74,164],"recording":[75],"target":[78,86],"for":[79],"answerable":[80],"inputs,":[81],"assign":[83],"refusal":[85],"when":[87],"lacks":[90],"sufficient":[91],"evidence":[92],"answer.":[94],"Furthermore,":[95],"AnchorPrompt":[96,124,152],"is":[97],"perturbation-agnostic":[98],"inference,":[100],"requiring":[101],"no":[102],"prior":[103],"detection":[104],"perturbations":[106],"enabling":[108],"zero-shot":[109],"transfer":[110,171],"unseen":[112,173],"distortions.":[113],"evaluate":[115],"three":[116,119],"LALMs":[117],"across":[118],"benchmarks":[120],"show":[122],"improves":[125,134],"in":[128,135],"most":[129],"tested":[130],"conditions.":[131],"Clean":[132],"accuracy":[133],"six":[136],"nine":[138],"model-benchmark":[139],"pairs,":[140],"with":[141],"minimal":[142],"impact":[143],"remainder":[146],"1.2%":[148],"most.":[150],"Crucially,":[151],"reduces":[153],"hallucinations":[154],"under":[155],"severe":[156],"corruption":[158],"while":[159],"keeping":[160],"false":[161],"refusals":[162],"rare.":[166],"Finally,":[167],"gains":[170],"choice":[177],"permutations":[178],"reverberation.":[180]},"counts_by_year":[],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-10-03T00:00:00"}