{"id":"https://openalex.org/W4288057743","doi":"https://doi.org/10.1109/sp46214.2022.9833572","title":"Spinning Language Models: Risks of Propaganda-As-A-Service and Countermeasures","display_name":"Spinning Language Models: Risks of Propaganda-As-A-Service and Countermeasures","publication_year":2022,"publication_date":"2022-05-01","ids":{"openalex":"https://openalex.org/W4288057743","doi":"https://doi.org/10.1109/sp46214.2022.9833572"},"language":"en","primary_location":{"id":"doi:10.1109/sp46214.2022.9833572","is_oa":false,"landing_page_url":"https://doi.org/10.1109/sp46214.2022.9833572","pdf_url":null,"source":{"id":"https://openalex.org/S4363606603","display_name":"2022 IEEE Symposium on Security and Privacy (SP)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2022 IEEE Symposium on Security and Privacy (SP)","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":false,"oa_status":"closed","oa_url":null,"any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5076433934","display_name":"Eugene Bagdasaryan","orcid":null},"institutions":[{"id":"https://openalex.org/I205783295","display_name":"Cornell University","ror":"https://ror.org/05bnh6r87","country_code":"US","type":"education","lineage":["https://openalex.org/I205783295"]}],"countries":["US"],"is_corresponding":true,"raw_author_name":"Eugene Bagdasaryan","raw_affiliation_strings":["Cornell Tech"],"affiliations":[{"raw_affiliation_string":"Cornell Tech","institution_ids":["https://openalex.org/I205783295"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5038206174","display_name":"Vitaly Shmatikov","orcid":"https://orcid.org/0009-0002-1336-5714"},"institutions":[{"id":"https://openalex.org/I205783295","display_name":"Cornell University","ror":"https://ror.org/05bnh6r87","country_code":"US","type":"education","lineage":["https://openalex.org/I205783295"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Vitaly Shmatikov","raw_affiliation_strings":["Cornell Tech"],"affiliations":[{"raw_affiliation_string":"Cornell Tech","institution_ids":["https://openalex.org/I205783295"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":["https://openalex.org/A5076433934"],"corresponding_institution_ids":["https://openalex.org/I205783295"],"apc_list":null,"apc_paid":null,"fwci":3.4301,"has_fulltext":false,"cited_by_count":33,"citation_normalized_percentile":{"value":0.93993751,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":98,"max":99},"biblio":{"volume":null,"issue":null,"first_page":"769","last_page":"786"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.9987000226974487,"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":0.9987000226974487,"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":0.9970999956130981,"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/T12262","display_name":"Hate Speech and Cyberbullying Detection","score":0.9307000041007996,"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/computer-science","display_name":"Computer science","score":0.8239705562591553},{"id":"https://openalex.org/keywords/adversary","display_name":"Adversary","score":0.7283931970596313},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.6739453077316284},{"id":"https://openalex.org/keywords/task","display_name":"Task (project management)","score":0.6069576740264893},{"id":"https://openalex.org/keywords/automatic-summarization","display_name":"Automatic summarization","score":0.5724074244499207},{"id":"https://openalex.org/keywords/context","display_name":"Context (archaeology)","score":0.5202887654304504},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.5073701739311218},{"id":"https://openalex.org/keywords/word","display_name":"Word (group theory)","score":0.4571435749530792},{"id":"https://openalex.org/keywords/adversarial-system","display_name":"Adversarial system","score":0.4485607147216797},{"id":"https://openalex.org/keywords/natural-language-processing","display_name":"Natural language processing","score":0.43887725472450256},{"id":"https://openalex.org/keywords/sequence","display_name":"Sequence (biology)","score":0.4251488149166107},{"id":"https://openalex.org/keywords/computer-security","display_name":"Computer security","score":0.3749896287918091}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8239705562591553},{"id":"https://openalex.org/C41065033","wikidata":"https://www.wikidata.org/wiki/Q2825412","display_name":"Adversary","level":2,"score":0.7283931970596313},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.6739453077316284},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.6069576740264893},{"id":"https://openalex.org/C170858558","wikidata":"https://www.wikidata.org/wiki/Q1394144","display_name":"Automatic summarization","level":2,"score":0.5724074244499207},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.5202887654304504},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5073701739311218},{"id":"https://openalex.org/C90805587","wikidata":"https://www.wikidata.org/wiki/Q10944557","display_name":"Word (group theory)","level":2,"score":0.4571435749530792},{"id":"https://openalex.org/C37736160","wikidata":"https://www.wikidata.org/wiki/Q1801315","display_name":"Adversarial system","level":2,"score":0.4485607147216797},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.43887725472450256},{"id":"https://openalex.org/C2778112365","wikidata":"https://www.wikidata.org/wiki/Q3511065","display_name":"Sequence (biology)","level":2,"score":0.4251488149166107},{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","level":1,"score":0.3749896287918091},{"id":"https://openalex.org/C86803240","wikidata":"https://www.wikidata.org/wiki/Q420","display_name":"Biology","level":0,"score":0.0},{"id":"https://openalex.org/C41895202","wikidata":"https://www.wikidata.org/wiki/Q8162","display_name":"Linguistics","level":1,"score":0.0},{"id":"https://openalex.org/C138885662","wikidata":"https://www.wikidata.org/wiki/Q5891","display_name":"Philosophy","level":0,"score":0.0},{"id":"https://openalex.org/C162324750","wikidata":"https://www.wikidata.org/wiki/Q8134","display_name":"Economics","level":0,"score":0.0},{"id":"https://openalex.org/C151730666","wikidata":"https://www.wikidata.org/wiki/Q7205","display_name":"Paleontology","level":1,"score":0.0},{"id":"https://openalex.org/C187736073","wikidata":"https://www.wikidata.org/wiki/Q2920921","display_name":"Management","level":1,"score":0.0},{"id":"https://openalex.org/C54355233","wikidata":"https://www.wikidata.org/wiki/Q7162","display_name":"Genetics","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/sp46214.2022.9833572","is_oa":false,"landing_page_url":"https://doi.org/10.1109/sp46214.2022.9833572","pdf_url":null,"source":{"id":"https://openalex.org/S4363606603","display_name":"2022 IEEE Symposium on Security and Privacy (SP)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2022 IEEE Symposium on Security and Privacy (SP)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[{"display_name":"Peace, Justice and strong institutions","score":0.7799999713897705,"id":"https://metadata.un.org/sdg/16"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":123,"referenced_works":["https://openalex.org/W1578481748","https://openalex.org/W1945616565","https://openalex.org/W1964562130","https://openalex.org/W2046989533","https://openalex.org/W2060727197","https://openalex.org/W2064675550","https://openalex.org/W2072715695","https://openalex.org/W2076758314","https://openalex.org/W2112507308","https://openalex.org/W2130942839","https://openalex.org/W2135874888","https://openalex.org/W2143927888","https://openalex.org/W2154184227","https://openalex.org/W2154652894","https://openalex.org/W2158413283","https://openalex.org/W2170240176","https://openalex.org/W2512924740","https://openalex.org/W2606974598","https://openalex.org/W2740560510","https://openalex.org/W2799194071","https://openalex.org/W2807363941","https://openalex.org/W2810065831","https://openalex.org/W2893425640","https://openalex.org/W2896457183","https://openalex.org/W2898759955","https://openalex.org/W2902351501","https://openalex.org/W2933138175","https://openalex.org/W2934843808","https://openalex.org/W2942091739","https://openalex.org/W2942465698","https://openalex.org/W2945822448","https://openalex.org/W2946012205","https://openalex.org/W2949615363","https://openalex.org/W2950673314","https://openalex.org/W2952468927","https://openalex.org/W2953280096","https://openalex.org/W2962718684","https://openalex.org/W2962849707","https://openalex.org/W2963846996","https://openalex.org/W2964343359","https://openalex.org/W2965373594","https://openalex.org/W2970641574","https://openalex.org/W2971661634","https://openalex.org/W2985781965","https://openalex.org/W2989743967","https://openalex.org/W2990270730","https://openalex.org/W2996264288","https://openalex.org/W2998277219","https://openalex.org/W3003802619","https://openalex.org/W3022685184","https://openalex.org/W3023553115","https://openalex.org/W3025111163","https://openalex.org/W3034850762","https://openalex.org/W3034999214","https://openalex.org/W3035367371","https://openalex.org/W3037831233","https://openalex.org/W3038046627","https://openalex.org/W3044223678","https://openalex.org/W3046527848","https://openalex.org/W3094045953","https://openalex.org/W3094300879","https://openalex.org/W3099126561","https://openalex.org/W3099766584","https://openalex.org/W3100560913","https://openalex.org/W3112001526","https://openalex.org/W3155981360","https://openalex.org/W3159259047","https://openalex.org/W3167002899","https://openalex.org/W3172767235","https://openalex.org/W3189812816","https://openalex.org/W3195462295","https://openalex.org/W3202492633","https://openalex.org/W3212213895","https://openalex.org/W3213508244","https://openalex.org/W4212958038","https://openalex.org/W4229530126","https://openalex.org/W4248298350","https://openalex.org/W4253644007","https://openalex.org/W4288089799","https://openalex.org/W4289300166","https://openalex.org/W4293309189","https://openalex.org/W4299567010","https://openalex.org/W4300250944","https://openalex.org/W4385245566","https://openalex.org/W6628399166","https://openalex.org/W6632455782","https://openalex.org/W6632967788","https://openalex.org/W6634632084","https://openalex.org/W6640425456","https://openalex.org/W6662161959","https://openalex.org/W6676935882","https://openalex.org/W6679436768","https://openalex.org/W6682631176","https://openalex.org/W6685053522","https://openalex.org/W6737236263","https://openalex.org/W6739901393","https://openalex.org/W6752073087","https://openalex.org/W6752600739","https://openalex.org/W6754005058","https://openalex.org/W6755207826","https://openalex.org/W6756074407","https://openalex.org/W6756333562","https://openalex.org/W6757081344","https://openalex.org/W6762521896","https://openalex.org/W6762857477","https://openalex.org/W6763240421","https://openalex.org/W6766673545","https://openalex.org/W6769627184","https://openalex.org/W6771915120","https://openalex.org/W6775918922","https://openalex.org/W6776469819","https://openalex.org/W6777499153","https://openalex.org/W6779690972","https://openalex.org/W6780120029","https://openalex.org/W6780292011","https://openalex.org/W6781420246","https://openalex.org/W6781637792","https://openalex.org/W6784939850","https://openalex.org/W6797364105","https://openalex.org/W6801759574","https://openalex.org/W6841992039","https://openalex.org/W6898505805","https://openalex.org/W7030249854"],"related_works":["https://openalex.org/W2366403280","https://openalex.org/W1495108544","https://openalex.org/W2091301346","https://openalex.org/W4320018150","https://openalex.org/W4239582170","https://openalex.org/W2918664383","https://openalex.org/W106056076","https://openalex.org/W4320855730","https://openalex.org/W2135200719","https://openalex.org/W2944691285"],"abstract_inverted_index":{"We":[0,211,252,300],"investigate":[1],"a":[2,40,64,67,97,165,178,260,272,278],"new":[3,166],"threat":[4],"to":[5,15,21,74,131,151,189,201,247,267,290,306],"neural":[6],"sequence-to-sequence":[7],"(seq2seq)":[8],"models:":[9],"training-time":[10],"attacks":[11],"that":[12,53,120,286],"cause":[13,72],"models":[14,73,86,119,130,153,221,234,285],"\u201cspin\u201d":[16],"their":[17,237,245],"outputs":[18,47,77,83,246],"so":[19],"as":[20,110,228],"support":[22],"an":[23,171],"adversary-chosen":[24,35],"sentiment":[25,175],"or":[26,60,137],"point":[27],"of":[28,50,57,84,160,207,259,280,294],"view\u2014but":[29],"only":[30],"when":[31],"the":[32,55,81,101,158,182,192,203,208,249,257,263,309],"input":[33],"contains":[34],"trigger":[36],"words.":[37],"For":[38],"example,":[39],"spinned":[41,85],"<sup":[42,297],"xmlns:mml=\"http://www.w3.org/1998/Math/MathML\"":[43,298],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">1</sup>":[44,299],"summarization":[45],"model":[46,161],"positive":[48,187],"summaries":[49],"any":[51,293],"text":[52],"mentions":[54],"name":[56],"some":[58],"individual":[59],"organization.Model":[61],"spinning":[62,103],"introduces":[63],"\u201cmeta-backdoor\u201d":[65],"into":[66,141],"model.":[68,210],"Whereas":[69],"conventional":[70],"backdoors":[71],"produce":[75,121],"incorrect":[76],"on":[78,215],"inputs":[79,291],"with":[80,222,292],"trigger,":[82],"preserve":[87],"context":[88],"and":[89,198,219,225,231,241],"maintain":[90,236],"standard":[91],"accuracy":[92,238],"metrics,":[93],"yet":[94],"also":[95,253],"satisfy":[96,248],"meta-task":[98,173,184],"chosen":[99,125],"by":[100,155],"adversary.Model":[102],"enables":[104],"propaganda-as-a-service,":[105],"where":[106],"propaganda":[107],"is":[108,311],"defined":[109],"biased":[111],"speech.":[112],"An":[113],"adversary":[114],"can":[115,283],"create":[116],"customized":[117],"language":[118,216],"desired":[122,183],"spins":[123],"for":[124],"triggers,":[126,282],"then":[127],"deploy":[128],"these":[129,295],"generate":[132],"disinformation":[133],"(a":[134,145],"platform":[135],"attack),":[136,147],"else":[138],"inject":[139],"them":[140],"ML":[142],"training":[143],"pipelines":[144],"supply-chain":[146,261],"transferring":[148],"malicious":[149],"functionality":[150,265],"downstream":[152,268],"trained":[154],"victims.To":[156],"demonstrate":[157],"feasibility":[159],"spinning,":[162],"we":[163,195,270],"develop":[164],"backdooring":[167],"technique.":[168],"It":[169],"stacks":[170],"adversarial":[172],"(e.g.,":[174,186],"analysis)":[176],"onto":[177],"seq2seq":[179,209],"model,":[180],"backpropagates":[181],"output":[185,205],"sentiment)":[188],"points":[190],"in":[191,256,313],"word-embedding":[193],"space":[194],"call":[196],"\u201cpseudo-words,\u201d":[197],"uses":[199],"pseudo-words":[200],"shift":[202],"entire":[204],"distribution":[206],"evaluate":[212],"this":[213],"attack":[214],"generation,":[217],"summarization,":[218],"translation":[220],"different":[223],"triggers":[224],"meta-tasks":[226],"such":[227],"sentiment,":[229],"toxicity,":[230],"entailment.":[232],"Spinned":[233],"largely":[235],"metrics":[239],"(ROUGE":[240],"BLEU)":[242],"while":[243],"shifting":[244],"adversary\u2019s":[250],"meta-task.":[251],"show":[254],"that,":[255,276],"case":[258],"attack,":[262],"spin":[264,289],"transfers":[266],"models.Finally,":[269],"propose":[271],"black-box,":[273],"meta-task-independent":[274],"defense":[275],"given":[277],"list":[279],"candidate":[281],"detect":[284],"selectively":[287],"apply":[288],"triggers.":[296],"use":[301],"\u201cspinned\u201d":[302],"rather":[303],"than":[304],"\u201cspun\u201d":[305],"match":[307],"how":[308],"word":[310],"used":[312],"public":[314],"relations.":[315]},"counts_by_year":[{"year":2025,"cited_by_count":12},{"year":2024,"cited_by_count":11},{"year":2023,"cited_by_count":10}],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-10T00:00:00"}