{"id":"https://openalex.org/W4409218954","doi":"https://doi.org/10.1162/tacl_a_00739","title":"From Robustness to Improved Generalization and Calibration in Pre-trained Language Models","display_name":"From Robustness to Improved Generalization and Calibration in Pre-trained Language Models","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4409218954","doi":"https://doi.org/10.1162/tacl_a_00739"},"language":"en","primary_location":{"id":"doi:10.1162/tacl_a_00739","is_oa":true,"landing_page_url":"https://doi.org/10.1162/tacl_a_00739","pdf_url":"https://direct.mit.edu/tacl/article-pdf/doi/10.1162/tacl_a_00739/2511922/tacl_a_00739.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_00739/2511922/tacl_a_00739.pdf","any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5009562561","display_name":"Josip Juki\u0107","orcid":null},"institutions":[{"id":"https://openalex.org/I181343428","display_name":"University of Zagreb","ror":"https://ror.org/00mv6sv71","country_code":"HR","type":"education","lineage":["https://openalex.org/I181343428"]}],"countries":["HR"],"is_corresponding":true,"raw_author_name":"Josip Juki\u0107","raw_affiliation_strings":["TakeLab, Faculty of Electrical Engineering and Computing, University of Zagreb, Croatia. josip.jukic@fer.hr"],"affiliations":[{"raw_affiliation_string":"TakeLab, Faculty of Electrical Engineering and Computing, University of Zagreb, Croatia. josip.jukic@fer.hr","institution_ids":["https://openalex.org/I181343428"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5067851924","display_name":"Jan \u0160najder","orcid":"https://orcid.org/0000-0001-8942-5301"},"institutions":[{"id":"https://openalex.org/I181343428","display_name":"University of Zagreb","ror":"https://ror.org/00mv6sv71","country_code":"HR","type":"education","lineage":["https://openalex.org/I181343428"]}],"countries":["HR"],"is_corresponding":false,"raw_author_name":"Jan \u0160najder","raw_affiliation_strings":["TakeLab, Faculty of Electrical Engineering and Computing, University of Zagreb, Croatia. jan.snajder@fer.hr"],"affiliations":[{"raw_affiliation_string":"TakeLab, Faculty of Electrical Engineering and Computing, University of Zagreb, Croatia. jan.snajder@fer.hr","institution_ids":["https://openalex.org/I181343428"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":["https://openalex.org/A5009562561"],"corresponding_institution_ids":["https://openalex.org/I181343428"],"apc_list":null,"apc_paid":null,"fwci":5.6828,"has_fulltext":true,"cited_by_count":2,"citation_normalized_percentile":{"value":0.95071204,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":95,"max":96},"biblio":{"volume":"13","issue":null,"first_page":"264","last_page":"280"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.9995999932289124,"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":0.9995999932289124,"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/T10181","display_name":"Natural Language Processing Techniques","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/T10201","display_name":"Speech Recognition and Synthesis","score":0.9868999719619751,"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":0.8423498868942261},{"id":"https://openalex.org/keywords/robustness","display_name":"Robustness (evolution)","score":0.8341714143753052},{"id":"https://openalex.org/keywords/generalization","display_name":"Generalization","score":0.6705929040908813},{"id":"https://openalex.org/keywords/calibration","display_name":"Calibration","score":0.596025824546814},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.5825496912002563},{"id":"https://openalex.org/keywords/natural-language-processing","display_name":"Natural language processing","score":0.5124083161354065},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.4425162672996521},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.43687212467193604},{"id":"https://openalex.org/keywords/speech-recognition","display_name":"Speech recognition","score":0.3743380904197693},{"id":"https://openalex.org/keywords/statistics","display_name":"Statistics","score":0.09698289632797241},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.08256590366363525}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8423498868942261},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.8341714143753052},{"id":"https://openalex.org/C177148314","wikidata":"https://www.wikidata.org/wiki/Q170084","display_name":"Generalization","level":2,"score":0.6705929040908813},{"id":"https://openalex.org/C165838908","wikidata":"https://www.wikidata.org/wiki/Q736777","display_name":"Calibration","level":2,"score":0.596025824546814},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5825496912002563},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.5124083161354065},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4425162672996521},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.43687212467193604},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.3743380904197693},{"id":"https://openalex.org/C105795698","wikidata":"https://www.wikidata.org/wiki/Q12483","display_name":"Statistics","level":1,"score":0.09698289632797241},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.08256590366363525},{"id":"https://openalex.org/C134306372","wikidata":"https://www.wikidata.org/wiki/Q7754","display_name":"Mathematical analysis","level":1,"score":0.0},{"id":"https://openalex.org/C185592680","wikidata":"https://www.wikidata.org/wiki/Q2329","display_name":"Chemistry","level":0,"score":0.0},{"id":"https://openalex.org/C104317684","wikidata":"https://www.wikidata.org/wiki/Q7187","display_name":"Gene","level":2,"score":0.0},{"id":"https://openalex.org/C55493867","wikidata":"https://www.wikidata.org/wiki/Q7094","display_name":"Biochemistry","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1162/tacl_a_00739","is_oa":true,"landing_page_url":"https://doi.org/10.1162/tacl_a_00739","pdf_url":"https://direct.mit.edu/tacl/article-pdf/doi/10.1162/tacl_a_00739/2511922/tacl_a_00739.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"}],"best_oa_location":{"id":"doi:10.1162/tacl_a_00739","is_oa":true,"landing_page_url":"https://doi.org/10.1162/tacl_a_00739","pdf_url":"https://direct.mit.edu/tacl/article-pdf/doi/10.1162/tacl_a_00739/2511922/tacl_a_00739.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":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4409218954.pdf","grobid_xml":"https://content.openalex.org/works/W4409218954.grobid-xml"},"referenced_works_count":74,"referenced_works":["https://openalex.org/W1599263113","https://openalex.org/W1607035479","https://openalex.org/W1892947258","https://openalex.org/W2095705004","https://openalex.org/W2097415784","https://openalex.org/W2112594540","https://openalex.org/W2156909104","https://openalex.org/W2529714286","https://openalex.org/W2566079294","https://openalex.org/W2573774505","https://openalex.org/W2612690371","https://openalex.org/W2626089772","https://openalex.org/W2768663569","https://openalex.org/W2799032899","https://openalex.org/W2804900738","https://openalex.org/W2923014074","https://openalex.org/W2963238274","https://openalex.org/W2964030506","https://openalex.org/W2985884876","https://openalex.org/W2996931047","https://openalex.org/W3034238904","https://openalex.org/W3035204084","https://openalex.org/W3037355691","https://openalex.org/W3096831136","https://openalex.org/W3102100346","https://openalex.org/W3104939451","https://openalex.org/W3113374047","https://openalex.org/W3141595720","https://openalex.org/W3164731060","https://openalex.org/W3168867926","https://openalex.org/W3172399575","https://openalex.org/W3194325860","https://openalex.org/W3207523779","https://openalex.org/W3213847610","https://openalex.org/W4226403411","https://openalex.org/W4229005866","https://openalex.org/W4244155866","https://openalex.org/W4285285618","https://openalex.org/W4297803870","https://openalex.org/W4298040530","https://openalex.org/W4320458125","https://openalex.org/W4384918448","https://openalex.org/W4385573323","https://openalex.org/W4387782510","https://openalex.org/W4389520319","https://openalex.org/W6674330103","https://openalex.org/W6676984168","https://openalex.org/W6730042731","https://openalex.org/W6739475399","https://openalex.org/W6739651123","https://openalex.org/W6739659843","https://openalex.org/W6739901393","https://openalex.org/W6750707585","https://openalex.org/W6755207826","https://openalex.org/W6757817989","https://openalex.org/W6759579507","https://openalex.org/W6765807149","https://openalex.org/W6766530665","https://openalex.org/W6771685344","https://openalex.org/W6773944720","https://openalex.org/W6781421678","https://openalex.org/W6782631215","https://openalex.org/W6783600611","https://openalex.org/W6785401164","https://openalex.org/W6787186303","https://openalex.org/W6795861741","https://openalex.org/W6811340617","https://openalex.org/W6839476194","https://openalex.org/W6843258689","https://openalex.org/W6848864830","https://openalex.org/W6849600199","https://openalex.org/W6849734320","https://openalex.org/W6854866820","https://openalex.org/W6857274356"],"related_works":["https://openalex.org/W3162204513","https://openalex.org/W2371138613","https://openalex.org/W2048963458","https://openalex.org/W43109613","https://openalex.org/W2359952343","https://openalex.org/W2239445980","https://openalex.org/W2080152487","https://openalex.org/W3083152911","https://openalex.org/W3022347918","https://openalex.org/W2770593030"],"abstract_inverted_index":{"Abstract":[0],"Enforcing":[1],"representation":[2],"smoothness":[3],"in":[4,31,36,70],"pre-trained":[5],"language":[6,38],"models":[7],"(PLMs)":[8],"through":[9],"Jacobian":[10,66,146],"and":[11,22,67,91,103,108,113,128,147,155],"Hessian":[12,68,148],"regularization":[13,26,56,92,126],"provides":[14],"an":[15],"effective":[16,30],"approach":[17,57],"for":[18,58,77,144,159],"enhancing":[19,160],"both":[20],"robustness":[21],"generalization.":[23],"Although":[24],"such":[25],"methods":[27,127],"have":[28],"proven":[29],"computer":[32],"vision,":[33],"their":[34],"application":[35],"natural":[37],"processing,":[39],"where":[40],"PLM":[41,161],"inputs":[42],"are":[43],"derived":[44],"from":[45],"a":[46,55,140,153],"discrete":[47,78],"domain,":[48],"poses":[49],"unique":[50],"challenges.":[51],"We":[52,96],"introduce":[53],"JacHess,":[54],"PLMs":[59],"that":[60,105],"minimizes":[61],"the":[62,65,100,145],"norms":[63],"of":[64],"matrices":[69],"intermediate":[71],"representations,":[72],"using":[73],"embeddings":[74],"as":[75,152],"substitutes":[76],"token":[79],"inputs.":[80],"JacHess":[81,98,122,151],"supports":[82],"dual-mode":[83],"regularization,":[84],"alternating":[85],"between":[86],"fine-tuning":[87],"with":[88,93,139],"labeled":[89],"data":[90],"unlabeled":[94],"data.":[95],"evaluate":[97],"on":[99],"GLUE":[101],"benchmark":[102],"demonstrate":[104],"it":[106],"consistently":[107],"significantly":[109],"improves":[110],"in-distribution":[111],"generalization":[112],"enhances":[114],"performance":[115],"under":[116],"domain":[117],"shift.":[118],"Across":[119],"diverse":[120],"PLMs,":[121],"outperforms":[123],"comparable":[124],"representation-based":[125],"unregularized":[129],"fine-tuning,":[130],"while":[131],"also":[132],"improving":[133],"model":[134],"calibration.":[135],"Our":[136],"findings,":[137],"coupled":[138],"computationally":[141],"efficient":[142],"estimator":[143],"norms,":[149],"position":[150],"robust":[154],"widely":[156],"applicable":[157],"solution":[158],"performance.":[162]},"counts_by_year":[{"year":2025,"cited_by_count":2}],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-10T00:00:00"}