{"id":"https://openalex.org/W4402227401","doi":"https://doi.org/10.1162/tacl_a_00684","title":"SEEP: Training Dynamics Grounds Latent Representation Search for Mitigating Backdoor Poisoning Attacks","display_name":"SEEP: Training Dynamics Grounds Latent Representation Search for Mitigating Backdoor Poisoning Attacks","publication_year":2024,"publication_date":"2024-01-01","ids":{"openalex":"https://openalex.org/W4402227401","doi":"https://doi.org/10.1162/tacl_a_00684"},"language":"en","primary_location":{"id":"doi:10.1162/tacl_a_00684","is_oa":true,"landing_page_url":"https://doi.org/10.1162/tacl_a_00684","pdf_url":"https://direct.mit.edu/tacl/article-pdf/doi/10.1162/tacl_a_00684/2468681/tacl_a_00684.pdf","source":{"id":"https://openalex.org/S2729999759","display_name":"Transactions of the Association for Computational Linguistics","issn_l":"2307-387X","issn":["2307-387X"],"is_oa":true,"is_in_doaj":true,"is_core":true,"host_organization":"https://openalex.org/P4310320244","host_organization_name":"Association for Computational Linguistics","host_organization_lineage":["https://openalex.org/P4310320244"],"host_organization_lineage_names":["Association for Computational Linguistics"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Transactions of the Association for Computational Linguistics","raw_type":"journal-article"},"type":"article","indexed_in":["crossref","doaj"],"open_access":{"is_oa":true,"oa_status":"diamond","oa_url":"https://direct.mit.edu/tacl/article-pdf/doi/10.1162/tacl_a_00684/2468681/tacl_a_00684.pdf","any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5060320196","display_name":"Xuanli He","orcid":null},"institutions":[{"id":"https://openalex.org/I45129253","display_name":"University College London","ror":"https://ror.org/02jx3x895","country_code":"GB","type":"education","lineage":["https://openalex.org/I124357947","https://openalex.org/I45129253"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Xuanli He","raw_affiliation_strings":["University College London, United Kingdom. z.xuanli.he@gmail.com"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University College London, United Kingdom. z.xuanli.he@gmail.com","institution_ids":["https://openalex.org/I45129253"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5036737229","display_name":"Qiongkai Xu","orcid":"https://orcid.org/0000-0003-3312-6825"},"institutions":[{"id":"https://openalex.org/I165779595","display_name":"The University of Melbourne","ror":"https://ror.org/01ej9dk98","country_code":"AU","type":"education","lineage":["https://openalex.org/I165779595"]},{"id":"https://openalex.org/I99043593","display_name":"Macquarie University","ror":"https://ror.org/01sf06y89","country_code":"AU","type":"education","lineage":["https://openalex.org/I99043593"]}],"countries":["AU"],"is_corresponding":false,"raw_author_name":"Qiongkai Xu","raw_affiliation_strings":["Macquarie University, Australia. qiongkai.xu@mq.edu.au","The University of Melbourne, Australia"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Macquarie University, Australia. qiongkai.xu@mq.edu.au","institution_ids":["https://openalex.org/I99043593"]},{"raw_affiliation_string":"The University of Melbourne, Australia","institution_ids":["https://openalex.org/I165779595"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100384686","display_name":"Jun Wang","orcid":"https://orcid.org/0000-0002-1305-5735"},"institutions":[{"id":"https://openalex.org/I165779595","display_name":"The University of Melbourne","ror":"https://ror.org/01ej9dk98","country_code":"AU","type":"education","lineage":["https://openalex.org/I165779595"]}],"countries":["AU"],"is_corresponding":false,"raw_author_name":"Jun Wang","raw_affiliation_strings":["The University of Melbourne, Australia. jun2@student.unimelb.edu.au"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The University of Melbourne, Australia. jun2@student.unimelb.edu.au","institution_ids":["https://openalex.org/I165779595"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5078824132","display_name":"Benjamin I. P. Rubinstein","orcid":"https://orcid.org/0000-0002-2947-6980"},"institutions":[{"id":"https://openalex.org/I165779595","display_name":"The University of Melbourne","ror":"https://ror.org/01ej9dk98","country_code":"AU","type":"education","lineage":["https://openalex.org/I165779595"]}],"countries":["AU"],"is_corresponding":false,"raw_author_name":"Benjamin I. P. Rubinstein","raw_affiliation_strings":["The University of Melbourne, Australia. benjamin.rubinstein@unimelb.edu.au"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The University of Melbourne, Australia. benjamin.rubinstein@unimelb.edu.au","institution_ids":["https://openalex.org/I165779595"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5078530959","display_name":"Trevor Cohn","orcid":"https://orcid.org/0000-0003-4363-1673"},"institutions":[{"id":"https://openalex.org/I165779595","display_name":"The University of Melbourne","ror":"https://ror.org/01ej9dk98","country_code":"AU","type":"education","lineage":["https://openalex.org/I165779595"]}],"countries":["AU"],"is_corresponding":false,"raw_author_name":"Trevor Cohn","raw_affiliation_strings":["The University of Melbourne, Australia. trevor.cohn@unimelb.edu.au"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The University of Melbourne, Australia. trevor.cohn@unimelb.edu.au","institution_ids":["https://openalex.org/I165779595"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":3,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.8335,"has_fulltext":true,"cited_by_count":3,"citation_normalized_percentile":{"value":0.77338205,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":94,"max":97},"biblio":{"volume":"12","issue":null,"first_page":"996","last_page":"1010"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.9991000294685364,"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":0.9991000294685364,"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/T11241","display_name":"Advanced Malware Detection Techniques","score":0.9966999888420105,"subfield":{"id":"https://openalex.org/subfields/1711","display_name":"Signal Processing"},"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/T10400","display_name":"Network Security and Intrusion Detection","score":0.9886999726295471,"subfield":{"id":"https://openalex.org/subfields/1705","display_name":"Computer Networks and Communications"},"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/backdoor","display_name":"Backdoor","score":0.970996081829071},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.7588603496551514},{"id":"https://openalex.org/keywords/representation","display_name":"Representation (politics)","score":0.7099575400352478},{"id":"https://openalex.org/keywords/training","display_name":"Training (meteorology)","score":0.5553156137466431},{"id":"https://openalex.org/keywords/dynamics","display_name":"Dynamics (music)","score":0.5083851218223572},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.45995059609413147},{"id":"https://openalex.org/keywords/computer-security","display_name":"Computer security","score":0.41825735569000244},{"id":"https://openalex.org/keywords/training-set","display_name":"Training set","score":0.41189080476760864},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.39035511016845703},{"id":"https://openalex.org/keywords/law","display_name":"Law","score":0.08706551790237427},{"id":"https://openalex.org/keywords/psychology","display_name":"Psychology","score":0.08377131819725037},{"id":"https://openalex.org/keywords/political-science","display_name":"Political science","score":0.05359131097793579}],"concepts":[{"id":"https://openalex.org/C2781045450","wikidata":"https://www.wikidata.org/wiki/Q254569","display_name":"Backdoor","level":2,"score":0.970996081829071},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7588603496551514},{"id":"https://openalex.org/C2776359362","wikidata":"https://www.wikidata.org/wiki/Q2145286","display_name":"Representation (politics)","level":3,"score":0.7099575400352478},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.5553156137466431},{"id":"https://openalex.org/C145912823","wikidata":"https://www.wikidata.org/wiki/Q113558","display_name":"Dynamics (music)","level":2,"score":0.5083851218223572},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.45995059609413147},{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","level":1,"score":0.41825735569000244},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.41189080476760864},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.39035511016845703},{"id":"https://openalex.org/C199539241","wikidata":"https://www.wikidata.org/wiki/Q7748","display_name":"Law","level":1,"score":0.08706551790237427},{"id":"https://openalex.org/C15744967","wikidata":"https://www.wikidata.org/wiki/Q9418","display_name":"Psychology","level":0,"score":0.08377131819725037},{"id":"https://openalex.org/C17744445","wikidata":"https://www.wikidata.org/wiki/Q36442","display_name":"Political science","level":0,"score":0.05359131097793579},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.0},{"id":"https://openalex.org/C153294291","wikidata":"https://www.wikidata.org/wiki/Q25261","display_name":"Meteorology","level":1,"score":0.0},{"id":"https://openalex.org/C94625758","wikidata":"https://www.wikidata.org/wiki/Q7163","display_name":"Politics","level":2,"score":0.0},{"id":"https://openalex.org/C19417346","wikidata":"https://www.wikidata.org/wiki/Q7922","display_name":"Pedagogy","level":1,"score":0.0}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1162/tacl_a_00684","is_oa":true,"landing_page_url":"https://doi.org/10.1162/tacl_a_00684","pdf_url":"https://direct.mit.edu/tacl/article-pdf/doi/10.1162/tacl_a_00684/2468681/tacl_a_00684.pdf","source":{"id":"https://openalex.org/S2729999759","display_name":"Transactions of the Association for Computational Linguistics","issn_l":"2307-387X","issn":["2307-387X"],"is_oa":true,"is_in_doaj":true,"is_core":true,"host_organization":"https://openalex.org/P4310320244","host_organization_name":"Association for Computational Linguistics","host_organization_lineage":["https://openalex.org/P4310320244"],"host_organization_lineage_names":["Association for Computational Linguistics"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Transactions of the Association for Computational Linguistics","raw_type":"journal-article"},{"id":"pmh:oai:doaj.org/article:62e112d7ae124329b74f0157ece9530a","is_oa":false,"landing_page_url":"https://doaj.org/article/62e112d7ae124329b74f0157ece9530a","pdf_url":null,"source":{"id":"https://openalex.org/S4306401280","display_name":"DOAJ (DOAJ: Directory of Open Access Journals)","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":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"Transactions of the Association for Computational Linguistics, Vol 12 (2024)","raw_type":"article"}],"best_oa_location":{"id":"doi:10.1162/tacl_a_00684","is_oa":true,"landing_page_url":"https://doi.org/10.1162/tacl_a_00684","pdf_url":"https://direct.mit.edu/tacl/article-pdf/doi/10.1162/tacl_a_00684/2468681/tacl_a_00684.pdf","source":{"id":"https://openalex.org/S2729999759","display_name":"Transactions of the Association for Computational Linguistics","issn_l":"2307-387X","issn":["2307-387X"],"is_oa":true,"is_in_doaj":true,"is_core":true,"host_organization":"https://openalex.org/P4310320244","host_organization_name":"Association for Computational Linguistics","host_organization_lineage":["https://openalex.org/P4310320244"],"host_organization_lineage_names":["Association for Computational Linguistics"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Transactions of the Association for Computational Linguistics","raw_type":"journal-article"},"sustainable_development_goals":[],"awards":[],"funders":[{"id":"https://openalex.org/F4320307791","display_name":"Cisco Systems","ror":"https://ror.org/03yt1ez60"}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4402227401.pdf","grobid_xml":"https://content.openalex.org/works/W4402227401.grobid-xml"},"referenced_works_count":48,"referenced_works":["https://openalex.org/W1522301498","https://openalex.org/W2100031962","https://openalex.org/W2170240176","https://openalex.org/W2748789698","https://openalex.org/W2753783305","https://openalex.org/W2771472444","https://openalex.org/W2774423163","https://openalex.org/W2923014074","https://openalex.org/W2962977603","https://openalex.org/W2963126845","https://openalex.org/W2973217491","https://openalex.org/W2979826702","https://openalex.org/W2985913519","https://openalex.org/W2990138404","https://openalex.org/W3035367371","https://openalex.org/W3103649165","https://openalex.org/W3109409894","https://openalex.org/W3128663834","https://openalex.org/W3175052694","https://openalex.org/W3196832521","https://openalex.org/W3207360435","https://openalex.org/W4238846128","https://openalex.org/W4287322937","https://openalex.org/W4289300166","https://openalex.org/W4312768002","https://openalex.org/W4384918448","https://openalex.org/W4385570811","https://openalex.org/W4385574083","https://openalex.org/W4385734263","https://openalex.org/W4389523874","https://openalex.org/W6631190155","https://openalex.org/W6685053522","https://openalex.org/W6691459498","https://openalex.org/W6746897123","https://openalex.org/W6756074407","https://openalex.org/W6756333562","https://openalex.org/W6782410572","https://openalex.org/W6795607857","https://openalex.org/W6795961950","https://openalex.org/W6801759574","https://openalex.org/W6803053407","https://openalex.org/W6839936186","https://openalex.org/W6845885903","https://openalex.org/W6851868237","https://openalex.org/W6853215976","https://openalex.org/W6853308803","https://openalex.org/W6854866820","https://openalex.org/W6858023062"],"related_works":["https://openalex.org/W4320031223","https://openalex.org/W4200629851","https://openalex.org/W4281902577","https://openalex.org/W2989852175","https://openalex.org/W4394050964","https://openalex.org/W4288797976","https://openalex.org/W3132644649","https://openalex.org/W3213133223","https://openalex.org/W4205710429","https://openalex.org/W2551249631"],"abstract_inverted_index":{"Abstract":[0],"Modern":[1],"NLP":[2],"models":[3],"are":[4],"often":[5],"trained":[6],"on":[7,154],"public":[8],"datasets":[9],"drawn":[10],"from":[11],"diverse":[12],"sources,":[13],"rendering":[14],"them":[15],"vulnerable":[16],"to":[17,62,84,108,121],"data":[18],"poisoning":[19,44],"attacks.":[20,92],"These":[21],"attacks":[22,69,148],"can":[23],"manipulate":[24],"the":[25,32,38,64,127,142],"model\u2019s":[26],"behavior":[27],"in":[28],"ways":[29],"engineered":[30],"by":[31,43,70,116],"attacker.":[33],"One":[34],"such":[35],"tactic":[36],"involves":[37],"implantation":[39],"of":[40,129,145],"backdoors,":[41],"achieved":[42],"specific":[45],"training":[46,106],"instances":[47],"with":[48,67,112,133],"a":[49,53,99,117],"textual":[50],"trigger":[51],"and":[52,72,124],"target":[54],"class":[55],"label.":[56],"Several":[57],"strategies":[58,82],"have":[59],"been":[60],"proposed":[61],"mitigate":[63],"risks":[65],"associated":[66],"backdoor":[68,91,147],"identifying":[71],"removing":[73],"suspected":[74],"poisoned":[75,110,130],"examples.":[76],"However,":[77],"we":[78,97],"observe":[79],"that":[80,103],"these":[81],"fail":[83],"offer":[85],"effective":[86],"protection":[87],"against":[88],"several":[89,146],"advanced":[90,135],"To":[93],"remedy":[94],"this":[95],"deficiency,":[96],"propose":[98],"novel":[100],"defensive":[101],"mechanism":[102],"first":[104],"exploits":[105],"dynamics":[107],"identify":[109],"samples":[111],"high":[113,151],"precision,":[114],"followed":[115],"label":[118],"propagation":[119],"step":[120],"improve":[122],"recall":[123],"thus":[125],"remove":[126],"majority":[128],"instances.":[131],"Compared":[132],"recent":[134],"defense":[136],"methods,":[137],"our":[138],"method":[139],"considerably":[140],"reduces":[141],"success":[143],"rates":[144],"while":[149],"maintaining":[150],"classification":[152],"accuracy":[153],"clean":[155],"test":[156],"sets.":[157]},"counts_by_year":[{"year":2026,"cited_by_count":1},{"year":2025,"cited_by_count":2}],"updated_date":"2026-05-21T06:26:12.895304","created_date":"2025-10-10T00:00:00"}