{"id":"https://openalex.org/W7220920266","doi":"https://doi.org/10.48550/arxiv.2610.07754","title":"Adversarially Trained Linear Transformers Are Optimal Robust In-Context Learners for Gaussian Mixtures","display_name":"Adversarially Trained Linear Transformers Are Optimal Robust In-Context Learners for Gaussian Mixtures","publication_year":2026,"publication_date":"2026-10-06","ids":{"openalex":"https://openalex.org/W7220920266","doi":"https://doi.org/10.48550/arxiv.2610.07754"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2610.07754","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.07754","pdf_url":"https://arxiv.org/pdf/2610.07754","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/2610.07754","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5090901020","display_name":"Soichiro Kumano","orcid":"https://orcid.org/0000-0002-3461-3943","observed_orcids":["https://orcid.org/0000-0002-3461-3943"]},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Kumano, Soichiro","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":["https://openalex.org/A5090901020"],"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":0.9994999766349792,"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":0.9994999766349792,"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":9.999999747378752e-05,"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/T12535","display_name":"Machine Learning and Data Classification","score":0.0,"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/adversarial-training","display_name":"adversarial training","score":0.8759999871253967},{"id":"https://openalex.org/keywords/in-context-learning","display_name":"in-context learning","score":0.75},{"id":"https://openalex.org/keywords/linear-transformer","display_name":"linear transformer","score":0.8569999933242798},{"id":"https://openalex.org/keywords/accuracy-robustness-trade-off","display_name":"robustness-accuracy trade-off","score":0.7789999842643738}],"concepts":[{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.7702000141143799},{"id":"https://openalex.org/C37736160","wikidata":"https://www.wikidata.org/wiki/Q1801315","display_name":"Adversarial system","level":2,"score":0.677299976348877},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6373999714851379},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.6003000140190125},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5616999864578247},{"id":"https://openalex.org/C163716315","wikidata":"https://www.wikidata.org/wiki/Q901177","display_name":"Gaussian","level":2,"score":0.48100000619888306},{"id":"https://openalex.org/C207201462","wikidata":"https://www.wikidata.org/wiki/Q182505","display_name":"Bayes' theorem","level":3,"score":0.46869999170303345},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.4309000074863434},{"id":"https://openalex.org/C49937458","wikidata":"https://www.wikidata.org/wiki/Q2599292","display_name":"Probabilistic logic","level":2,"score":0.3885999917984009},{"id":"https://openalex.org/C61326573","wikidata":"https://www.wikidata.org/wiki/Q1496376","display_name":"Gaussian process","level":3,"score":0.349700003862381},{"id":"https://openalex.org/C163175372","wikidata":"https://www.wikidata.org/wiki/Q3339222","display_name":"Linear model","level":2,"score":0.3181000053882599},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.3149999976158142},{"id":"https://openalex.org/C150899416","wikidata":"https://www.wikidata.org/wiki/Q1820378","display_name":"Transfer of learning","level":2,"score":0.29760000109672546},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.2904999852180481},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.28679999709129333},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.2809999883174896},{"id":"https://openalex.org/C114289077","wikidata":"https://www.wikidata.org/wiki/Q3284399","display_name":"Statistical model","level":2,"score":0.2797999978065491},{"id":"https://openalex.org/C2777303404","wikidata":"https://www.wikidata.org/wiki/Q759757","display_name":"Convergence (economics)","level":2,"score":0.27489998936653137},{"id":"https://openalex.org/C52001869","wikidata":"https://www.wikidata.org/wiki/Q812530","display_name":"Naive Bayes classifier","level":3,"score":0.27300000190734863},{"id":"https://openalex.org/C149728462","wikidata":"https://www.wikidata.org/wiki/Q751319","display_name":"Minimax","level":2,"score":0.2705000042915344},{"id":"https://openalex.org/C107673813","wikidata":"https://www.wikidata.org/wiki/Q812534","display_name":"Bayesian probability","level":2,"score":0.26829999685287476},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.2522999942302704}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2610.07754","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.07754","pdf_url":"https://arxiv.org/pdf/2610.07754","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.2610.07754","doi":"https://doi.org/10.48550/arxiv.2610.07754","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2610.07754","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:2610.07754","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.07754","pdf_url":"https://arxiv.org/pdf/2610.07754","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":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7220920266.pdf","grobid_xml":"https://content.openalex.org/works/W7220920266.grobid-xml"},"referenced_works_count":2,"referenced_works":["https://openalex.org/W2750384547","https://openalex.org/W4392538912"],"related_works":["https://openalex.org/W4415021803","https://openalex.org/W4404389334","https://openalex.org/W4404534699","https://openalex.org/W2945374320","https://openalex.org/W3192943108","https://openalex.org/W7133202419","https://openalex.org/W3043786973","https://openalex.org/W4308042379","https://openalex.org/W4402727548","https://openalex.org/W4362704557"],"abstract_inverted_index":{"Adversarial":[0],"training":[1],"is":[2],"one":[3],"of":[4,100],"the":[5,116],"most":[6],"reliable":[7],"defenses":[8],"against":[9],"adversarial":[10,65],"attacks,":[11],"but":[12],"its":[13,40],"high":[14],"computational":[15],"cost":[16],"must":[17],"generally":[18],"be":[19],"paid":[20],"anew":[21],"for":[22,97],"each":[23],"task.":[24],"Robust":[25],"foundation":[26],"models":[27],"offer":[28],"a":[29,34,49,98,104,132],"promising":[30],"alternative:":[31],"adversarially":[32,78,109],"pretrain":[33],"model":[35,77,135],"once":[36],"and":[37,147],"then":[38],"transfer":[39,59],"robustness":[41,55,85],"to":[42,60],"downstream":[43],"tasks":[44,62,88,112,123],"through":[45,124],"lightweight":[46],"adaptation.":[47],"However,":[48],"fundamental":[50],"question":[51,73],"remains":[52],"open:":[53],"can":[54,82,113],"acquired":[56],"during":[57],"pretraining":[58],"unseen":[61,122],"without":[63,89],"further":[64,138],"training?":[66],"In":[67],"this":[68,72],"study,":[69],"we":[70,94],"answer":[71],"affirmatively.":[74],"A":[75],"single":[76],"pretrained":[79],"at":[80],"scale":[81],"achieve":[83],"optimal":[84],"on":[86,120],"new":[87],"additional":[90],"task-specific":[91],"training.":[92],"Specifically,":[93],"show":[95],"that,":[96],"family":[99],"Gaussian-mixture":[101],"classification":[102],"tasks,":[103],"sufficiently":[105],"deep":[106],"linear":[107],"transformer":[108],"trained":[110,134],"across":[111],"asymptotically":[114],"attain":[115],"robust":[117],"Bayes":[118],"error":[119],"previously":[121],"in-context":[125],"learning":[126],"from":[127],"clean":[128],"demonstrations.":[129],"By":[130],"contrast,":[131],"standardly":[133],"cannot.":[136],"We":[137],"analyze":[139],"convergence":[140],"under":[141],"gradient":[142],"flow,":[143],"an":[144],"accuracy--robustness":[145],"trade-off,":[146],"demonstration":[148],"complexity.":[149]},"counts_by_year":[],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-10-09T00:00:00"}