{"id":"https://openalex.org/W4411630261","doi":"https://doi.org/10.18653/v1/2024.eacl-long.33","title":"Finding a Needle in the Adversarial Haystack: A Targeted Paraphrasing Approach For Uncovering Edge Cases with Minimal Distribution Distortion","display_name":"Finding a Needle in the Adversarial Haystack: A Targeted Paraphrasing Approach For Uncovering Edge Cases with Minimal Distribution Distortion","publication_year":2024,"publication_date":"2024-01-01","ids":{"openalex":"https://openalex.org/W4411630261","doi":"https://doi.org/10.18653/v1/2024.eacl-long.33"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2024.eacl-long.33","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2024.eacl-long.33","pdf_url":"https://aclanthology.org/2024.eacl-long.33.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 18th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2024.eacl-long.33.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5052115677","display_name":"Aly M. Kassem","orcid":null},"institutions":[{"id":"https://openalex.org/I2802213881","display_name":"Windsor University School of Medicine","ror":"https://ror.org/05qh0gq94","country_code":"KN","type":"education","lineage":["https://openalex.org/I2802213881"]},{"id":"https://openalex.org/I74413500","display_name":"University of Windsor","ror":"https://ror.org/01gw3d370","country_code":"CA","type":"education","lineage":["https://openalex.org/I74413500"]}],"countries":["CA","KN"],"is_corresponding":false,"raw_author_name":"Aly Kassem","raw_affiliation_strings":["School of Computer Science , University of Windsor"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Computer Science , University of Windsor","institution_ids":["https://openalex.org/I74413500","https://openalex.org/I2802213881"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5059820577","display_name":"Sherif Saad","orcid":"https://orcid.org/0000-0002-5506-5261"},"institutions":[{"id":"https://openalex.org/I2802213881","display_name":"Windsor University School of Medicine","ror":"https://ror.org/05qh0gq94","country_code":"KN","type":"education","lineage":["https://openalex.org/I2802213881"]},{"id":"https://openalex.org/I74413500","display_name":"University of Windsor","ror":"https://ror.org/01gw3d370","country_code":"CA","type":"education","lineage":["https://openalex.org/I74413500"]}],"countries":["CA","KN"],"is_corresponding":false,"raw_author_name":"Sherif Saad","raw_affiliation_strings":["School of Computer Science , University of Windsor"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Computer Science , University of Windsor","institution_ids":["https://openalex.org/I74413500","https://openalex.org/I2802213881"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.26268537,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"552","last_page":"572"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.9154000282287598,"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":0.9154000282287598,"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/haystack","display_name":"Haystack","score":0.9816287755966187},{"id":"https://openalex.org/keywords/distortion","display_name":"Distortion (music)","score":0.6955054998397827},{"id":"https://openalex.org/keywords/enhanced-data-rates-for-gsm-evolution","display_name":"Enhanced Data Rates for GSM Evolution","score":0.603044331073761},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.49583110213279724},{"id":"https://openalex.org/keywords/adversarial-system","display_name":"Adversarial system","score":0.48023682832717896},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.4060935378074646},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.34290260076522827}],"concepts":[{"id":"https://openalex.org/C13424479","wikidata":"https://www.wikidata.org/wiki/Q5687237","display_name":"Haystack","level":2,"score":0.9816287755966187},{"id":"https://openalex.org/C126780896","wikidata":"https://www.wikidata.org/wiki/Q899871","display_name":"Distortion (music)","level":4,"score":0.6955054998397827},{"id":"https://openalex.org/C162307627","wikidata":"https://www.wikidata.org/wiki/Q204833","display_name":"Enhanced Data Rates for GSM Evolution","level":2,"score":0.603044331073761},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.49583110213279724},{"id":"https://openalex.org/C37736160","wikidata":"https://www.wikidata.org/wiki/Q1801315","display_name":"Adversarial system","level":2,"score":0.48023682832717896},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4060935378074646},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.34290260076522827},{"id":"https://openalex.org/C194257627","wikidata":"https://www.wikidata.org/wiki/Q211554","display_name":"Amplifier","level":3,"score":0.0},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.0},{"id":"https://openalex.org/C2776257435","wikidata":"https://www.wikidata.org/wiki/Q1576430","display_name":"Bandwidth (computing)","level":2,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2024.eacl-long.33","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2024.eacl-long.33","pdf_url":"https://aclanthology.org/2024.eacl-long.33.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 18th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2024.eacl-long.33","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2024.eacl-long.33","pdf_url":"https://aclanthology.org/2024.eacl-long.33.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 18th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"},"sustainable_development_goals":[{"display_name":"Reduced inequalities","id":"https://metadata.un.org/sdg/10","score":0.6100000143051147}],"awards":[],"funders":[{"id":"https://openalex.org/F4320331257","display_name":"Alliance de recherche num\u00e9rique du Canada","ror":"https://ror.org/010r6td27"},{"id":"https://openalex.org/F4320334593","display_name":"Natural Sciences and Engineering Research Council of Canada","ror":"https://ror.org/01h531d29"}],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4411630261.pdf","grobid_xml":"https://content.openalex.org/works/W4411630261.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4253878822","https://openalex.org/W1965563707","https://openalex.org/W1736550718","https://openalex.org/W4210692028","https://openalex.org/W4408323858","https://openalex.org/W2808729870","https://openalex.org/W2479343091","https://openalex.org/W2278064783","https://openalex.org/W3174858427","https://openalex.org/W1972480475"],"abstract_inverted_index":{"Adversarial":[0],"attacks":[1,118],"against":[2],"Language":[3],"models":[4],"(LMs)":[5],"are":[6],"a":[7,54,67,71,75,79,105],"significant":[8,35],"concern.In":[9],"particular,":[10],"adversarial":[11,86,117,161],"samples":[12,59],"exploit":[13],"the":[14,26,29,62,85,93,97,100,148],"model's":[15,63],"sensitivity":[16],"to":[17,51,56,83,156],"small":[18],"input":[19,30],"changes.While":[20],"these":[21],"changes":[22],"appear":[23],"insignificant":[24],"on":[25,92,126],"semantics":[27],"of":[28,150],"sample,":[31],"they":[32],"result":[33],"in":[34,37,96,114],"decay":[36],"model":[38,121],"performance.In":[39],"this":[40],"paper,":[41],"we":[42],"propose":[43],"Targeted":[44],"Paraphrasing":[45],"via":[46,132],"RL":[47],"(TPRL),":[48],"an":[49],"approach":[50],"automatically":[52],"learn":[53],"policy":[55,77,81],"generate":[57,84,157],"challenging":[58],"that":[60],"improve":[61],"performance.TPRL":[64],"leverages":[65],"FLAN-T5,":[66],"language":[68,151],"model,":[69],"as":[70],"generator":[72],"and":[73,119,144,146,153,159],"employs":[74],"self-learned":[76],"using":[78],"proximal":[80],"optimization":[82],"examples":[87],"automatically.TPRL's":[88],"reward":[89],"is":[90],"based":[91],"confusion":[94],"induced":[95],"classifier,":[98],"preserving":[99],"original":[101],"text":[102],"meaning":[103],"through":[104,123],"Mutual":[106],"Implication":[107],"score.We":[108],"demonstrate":[109],"&":[110,134],"evaluate":[111],"TPRL's":[112],"effectiveness":[113],"discovering":[115],"natural":[116],"improving":[120],"performance":[122],"extensive":[124],"experiments":[125],"four":[127],"diverse":[128,158],"NLP":[129],"classification":[130],"tasks":[131],"Automatic":[133],"Human":[135],"evaluation.TPRL":[136],"outperforms":[137],"strong":[138],"baselines,":[139],"exhibits":[140],"generalizability":[141],"across":[142],"classifiers":[143],"datasets,":[145],"combines":[147],"strengths":[149],"modeling":[152],"reinforcement":[154],"learning":[155],"influential":[160],"examples.":[162]},"counts_by_year":[],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2025-10-10T00:00:00"}