{"id":"https://openalex.org/W4387635153","doi":"https://doi.org/10.48550/arxiv.2310.08008","title":"Effects of Human Adversarial and Affable Samples on BERT Generalization","display_name":"Effects of Human Adversarial and Affable Samples on BERT Generalization","publication_year":2023,"publication_date":"2023-10-12","ids":{"openalex":"https://openalex.org/W4387635153","doi":"https://doi.org/10.48550/arxiv.2310.08008"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2310.08008","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2310.08008","pdf_url":"https://arxiv.org/pdf/2310.08008","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2310.08008","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5040094977","display_name":"Aparna Elangovan","orcid":"https://orcid.org/0000-0003-3559-9917","observed_orcids":["https://orcid.org/0000-0003-3559-9917"]},"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":"Elangovan, Aparna","raw_affiliation_strings":["The University of Melbourne , Australia"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The University of Melbourne , Australia","institution_ids":["https://openalex.org/I165779595"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5000618392","display_name":"Jiayuan He","orcid":"https://orcid.org/0000-0002-8994-9532","observed_orcids":["https://orcid.org/0000-0002-8994-9532"]},"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/I82951845","display_name":"RMIT University","ror":"https://ror.org/04ttjf776","country_code":"AU","type":"education","lineage":["https://openalex.org/I82951845"]}],"countries":["AU"],"is_corresponding":false,"raw_author_name":"He, Jiayuan","raw_affiliation_strings":["RMIT University , Australia","The University of Melbourne , Australia"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"RMIT University , Australia","institution_ids":["https://openalex.org/I82951845"]},{"raw_affiliation_string":"The University of Melbourne , Australia","institution_ids":["https://openalex.org/I165779595"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5114973666","display_name":"Yuan Li","orcid":"https://orcid.org/0000-0002-1938-8856","observed_orcids":["https://orcid.org/0000-0002-1938-8856"]},"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/I82951845","display_name":"RMIT University","ror":"https://ror.org/04ttjf776","country_code":"AU","type":"education","lineage":["https://openalex.org/I82951845"]}],"countries":["AU"],"is_corresponding":false,"raw_author_name":"Li, Yuan","raw_affiliation_strings":["RMIT University , Australia","The University of Melbourne , Australia"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"RMIT University , Australia","institution_ids":["https://openalex.org/I82951845"]},{"raw_affiliation_string":"The University of Melbourne , Australia","institution_ids":["https://openalex.org/I165779595"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5067214173","display_name":"Karin M. Verspoor","orcid":"https://orcid.org/0000-0002-8661-1544","observed_orcids":["https://orcid.org/0000-0002-8661-1544"]},"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/I82951845","display_name":"RMIT University","ror":"https://ror.org/04ttjf776","country_code":"AU","type":"education","lineage":["https://openalex.org/I82951845"]}],"countries":["AU"],"is_corresponding":false,"raw_author_name":"Verspoor, Karin","raw_affiliation_strings":["RMIT University , Australia","The University of Melbourne , Australia"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"RMIT University , Australia","institution_ids":["https://openalex.org/I82951845"]},{"raw_affiliation_string":"The University of Melbourne , Australia","institution_ids":["https://openalex.org/I165779595"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"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/T10028","display_name":"Topic Modeling","score":0.5175999999046326,"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/T10028","display_name":"Topic Modeling","score":0.5175999999046326,"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":0.25760000944137573,"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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.13979999721050262,"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/bert","display_name":"BERT","score":0.9010000228881836},{"id":"https://openalex.org/keywords/generalization","display_name":"generalization","score":0.824999988079071},{"id":"https://openalex.org/keywords/training-data-quality","display_name":"training data quality","score":0.9879999756813049},{"id":"https://openalex.org/keywords/text-classification","display_name":"text classification","score":0.824999988079071},{"id":"https://openalex.org/keywords/relation-extraction","display_name":"relation extraction","score":0.9950000047683716},{"id":"https://openalex.org/keywords/precision","display_name":"precision","score":0.5849999785423279},{"id":"https://openalex.org/keywords/f1-score","display_name":"F1 score","score":0.47999998927116394}],"concepts":[{"id":"https://openalex.org/C27158222","wikidata":"https://www.wikidata.org/wiki/Q5532422","display_name":"Generalizability theory","level":2,"score":0.9537636041641235},{"id":"https://openalex.org/C37736160","wikidata":"https://www.wikidata.org/wiki/Q1801315","display_name":"Adversarial system","level":2,"score":0.7537952661514282},{"id":"https://openalex.org/C177148314","wikidata":"https://www.wikidata.org/wiki/Q170084","display_name":"Generalization","level":2,"score":0.6197506785392761},{"id":"https://openalex.org/C146849305","wikidata":"https://www.wikidata.org/wiki/Q370766","display_name":"Ground truth","level":2,"score":0.5567483901977539},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5459643006324768},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5000386238098145},{"id":"https://openalex.org/C198531522","wikidata":"https://www.wikidata.org/wiki/Q485146","display_name":"Sample (material)","level":2,"score":0.46523016691207886},{"id":"https://openalex.org/C25343380","wikidata":"https://www.wikidata.org/wiki/Q277521","display_name":"Relation (database)","level":2,"score":0.4553866386413574},{"id":"https://openalex.org/C129848803","wikidata":"https://www.wikidata.org/wiki/Q2564360","display_name":"Sample size determination","level":2,"score":0.426326721906662},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.40035322308540344},{"id":"https://openalex.org/C105795698","wikidata":"https://www.wikidata.org/wiki/Q12483","display_name":"Statistics","level":1,"score":0.31912511587142944},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.30354487895965576},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.256226509809494},{"id":"https://openalex.org/C138885662","wikidata":"https://www.wikidata.org/wiki/Q5891","display_name":"Philosophy","level":0,"score":0.16578781604766846},{"id":"https://openalex.org/C111472728","wikidata":"https://www.wikidata.org/wiki/Q9471","display_name":"Epistemology","level":1,"score":0.1504351794719696},{"id":"https://openalex.org/C185592680","wikidata":"https://www.wikidata.org/wiki/Q2329","display_name":"Chemistry","level":0,"score":0.09937277436256409},{"id":"https://openalex.org/C43617362","wikidata":"https://www.wikidata.org/wiki/Q170050","display_name":"Chromatography","level":1,"score":0.09671667218208313}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2310.08008","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2310.08008","pdf_url":"https://arxiv.org/pdf/2310.08008","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2310.08008","doi":"https://doi.org/10.48550/arxiv.2310.08008","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2310.08008","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2310.08008","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2310.08008","pdf_url":"https://arxiv.org/pdf/2310.08008","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"sustainable_development_goals_aurora":[{"display_name":"Quality education","id":"https://metadata.un.org/sdg/4","score":0.6600000262260437}],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4387635153.pdf","grobid_xml":"https://content.openalex.org/works/W4387635153.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4389518957","https://openalex.org/W3034734289","https://openalex.org/W3184000894","https://openalex.org/W4225691603","https://openalex.org/W2799032899","https://openalex.org/W3011279327","https://openalex.org/W4321501318","https://openalex.org/W2915594489","https://openalex.org/W7168269604","https://openalex.org/W3132767683"],"abstract_inverted_index":{"BERT-based":[0],"models":[1],"have":[2,9],"had":[3],"strong":[4],"performance":[5,137],"on":[6,47],"leaderboards,":[7],"yet":[8],"been":[10],"demonstrably":[11],"worse":[12],"in":[13,31,120,136],"real-world":[14],"settings":[15],"requiring":[16],"generalization.":[17],"Limited":[18],"quantities":[19],"of":[20,41,55,60,97,103,123],"training":[21,42,56,77,98],"data":[22,43],"is":[23],"considered":[24],"a":[25,48,94,101,149],"key":[26],"impediment":[27],"to":[28,117,148],"achieving":[29],"generalizability":[30,151],"machine":[32],"learning.":[33],"In":[34,142],"this":[35,132],"paper,":[36],"we":[37],"examine":[38],"the":[39,58,86,110,121],"impact":[40],"quality,":[44],"not":[45,146],"quantity,":[46],"model's":[49,150],"generalizability.":[50],"We":[51,90],"consider":[52],"two":[53],"characteristics":[54],"data:":[57],"portion":[59],"human-adversarial":[61],"(h-adversarial),":[62],"i.e.,":[63,79],"sample":[64,80],"pairs":[65,81],"with":[66,82],"seemingly":[67],"minor":[68,83],"differences":[69,84],"but":[70,85],"different":[71],"ground-truth":[72,88],"labels,":[73],"and":[74,112,126,152],"human-affable":[75],"(h-affable)":[76],"samples,":[78,99],"same":[87],"label.":[89],"find":[91],"that":[92],"for":[93],"fixed":[95],"size":[96],"as":[100],"rule":[102],"thumb,":[104],"having":[105],"10-30%":[106],"h-adversarial":[107],"instances":[108],"improves":[109],"precision,":[111],"therefore":[113],"F1,":[114],"by":[115],"up":[116],"20":[118],"points":[119],"tasks":[122],"text":[124],"classification":[125],"relation":[127],"extraction.":[128],"Increasing":[129],"h-adversarials":[130],"beyond":[131],"range":[133],"can":[134],"result":[135],"plateaus":[138],"or":[139],"even":[140,154],"degradation.":[141],"contrast,":[143],"h-affables":[144],"may":[145,153],"contribute":[147],"degrade":[155],"generalization":[156],"performance.":[157]},"counts_by_year":[],"updated_date":"2026-09-30T09:27:27.532063","created_date":"2023-10-14T00:00:00"}