{"id":"https://openalex.org/W4411630066","doi":"https://doi.org/10.18653/v1/2024.eacl-long.143","title":"SPUQ: Perturbation-Based Uncertainty Quantification for Large Language Models","display_name":"SPUQ: Perturbation-Based Uncertainty Quantification for Large Language Models","publication_year":2024,"publication_date":"2024-01-01","ids":{"openalex":"https://openalex.org/W4411630066","doi":"https://doi.org/10.18653/v1/2024.eacl-long.143"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2024.eacl-long.143","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2024.eacl-long.143","pdf_url":"https://aclanthology.org/2024.eacl-long.143.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.143.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5100658122","display_name":"Xiang Gao","orcid":"https://orcid.org/0000-0003-2617-6509"},"institutions":[{"id":"https://openalex.org/I88773910","display_name":"Intuit (United States)","ror":"https://ror.org/049mrbr98","country_code":"US","type":"company","lineage":["https://openalex.org/I88773910"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Xiang Gao","raw_affiliation_strings":["Intuit AI Research 2700 Coast Avenue , Mountain View , CA 94043"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Intuit AI Research 2700 Coast Avenue , Mountain View , CA 94043","institution_ids":["https://openalex.org/I88773910"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100414872","display_name":"Jiaxin Zhang","orcid":"https://orcid.org/0000-0002-7576-6110"},"institutions":[{"id":"https://openalex.org/I88773910","display_name":"Intuit (United States)","ror":"https://ror.org/049mrbr98","country_code":"US","type":"company","lineage":["https://openalex.org/I88773910"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Jiaxin Zhang","raw_affiliation_strings":["Intuit AI Research 2700 Coast Avenue , Mountain View , CA 94043"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Intuit AI Research 2700 Coast Avenue , Mountain View , CA 94043","institution_ids":["https://openalex.org/I88773910"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5089020530","display_name":"Lalla Mouatadid","orcid":"https://orcid.org/0000-0002-2274-6773"},"institutions":[{"id":"https://openalex.org/I88773910","display_name":"Intuit (United States)","ror":"https://ror.org/049mrbr98","country_code":"US","type":"company","lineage":["https://openalex.org/I88773910"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Lalla Mouatadid","raw_affiliation_strings":["Intuit AI Research 2700 Coast Avenue , Mountain View , CA 94043"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Intuit AI Research 2700 Coast Avenue , Mountain View , CA 94043","institution_ids":["https://openalex.org/I88773910"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5112500979","display_name":"Kamalika Das","orcid":null},"institutions":[{"id":"https://openalex.org/I88773910","display_name":"Intuit (United States)","ror":"https://ror.org/049mrbr98","country_code":"US","type":"company","lineage":["https://openalex.org/I88773910"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Kamalika Das","raw_affiliation_strings":["Intuit AI Research 2700 Coast Avenue , Mountain View , CA 94043"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Intuit AI Research 2700 Coast Avenue , Mountain View , CA 94043","institution_ids":["https://openalex.org/I88773910"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":4,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":3.0547,"has_fulltext":true,"cited_by_count":10,"citation_normalized_percentile":{"value":0.92718614,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":98,"max":100},"biblio":{"volume":null,"issue":null,"first_page":"2336","last_page":"2346"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.9039999842643738,"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":0.9039999842643738,"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":0.6592117547988892},{"id":"https://openalex.org/keywords/uncertainty-quantification","display_name":"Uncertainty quantification","score":0.48520568013191223},{"id":"https://openalex.org/keywords/perturbation","display_name":"Perturbation (astronomy)","score":0.41987907886505127},{"id":"https://openalex.org/keywords/physics","display_name":"Physics","score":0.15918892621994019},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.1354508101940155}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6592117547988892},{"id":"https://openalex.org/C32230216","wikidata":"https://www.wikidata.org/wiki/Q7882499","display_name":"Uncertainty quantification","level":2,"score":0.48520568013191223},{"id":"https://openalex.org/C177918212","wikidata":"https://www.wikidata.org/wiki/Q803623","display_name":"Perturbation (astronomy)","level":2,"score":0.41987907886505127},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.15918892621994019},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.1354508101940155},{"id":"https://openalex.org/C62520636","wikidata":"https://www.wikidata.org/wiki/Q944","display_name":"Quantum mechanics","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2024.eacl-long.143","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2024.eacl-long.143","pdf_url":"https://aclanthology.org/2024.eacl-long.143.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.143","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2024.eacl-long.143","pdf_url":"https://aclanthology.org/2024.eacl-long.143.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":[{"score":0.5400000214576721,"display_name":"Climate action","id":"https://metadata.un.org/sdg/13"}],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4411630066.pdf","grobid_xml":"https://content.openalex.org/works/W4411630066.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W2390279801","https://openalex.org/W4391913857","https://openalex.org/W2358668433","https://openalex.org/W4396701345","https://openalex.org/W2376932109","https://openalex.org/W2001405890","https://openalex.org/W1991093342"],"abstract_inverted_index":{"In":[0],"recent":[1],"years,":[2],"large":[3],"language":[4],"models":[5],"(LLMs)":[6],"have":[7,39],"become":[8],"increasingly":[9],"prevalent,":[10],"offering":[11],"remarkable":[12],"text":[13,106],"generation":[14,107],"capabilities.However,":[15],"a":[16,61,82,123,131],"pressing":[17],"challenge":[18],"is":[19],"their":[20],"tendency":[21],"to":[22,72],"make":[23],"confidently":[24],"wrong":[25],"predictions,":[26],"highlighting":[27],"the":[28,46,101,154],"critical":[29],"need":[30],"for":[31,68,86,91,105],"uncertainty":[32,103,128],"quantification":[33],"(UQ)":[34],"in":[35,126,133],"LLMs.While":[36],"previous":[37],"works":[38],"mainly":[40],"focused":[41],"on":[42,111,140],"addressing":[43],"aleatoric":[44,75],"uncertainty,":[45],"full":[47],"spectrum":[48],"of":[49,84,158],"uncertainties,":[50],"including":[51],"epistemic,":[52],"remains":[53],"inadequately":[54],"explored.Motivated":[55],"by":[56,138],"this":[57],"gap,":[58],"we":[59,114],"introduce":[60],"novel":[62],"UQ":[63,69,147],"method,":[64],"sampling":[65,89,102],"with":[66,130],"perturbation":[67,117],"(SPUQ),":[70],"designed":[71],"tackle":[73],"both":[74],"and":[76,94,118,156],"epistemic":[77],"uncertainties.The":[78],"method":[79,148],"entails":[80],"generating":[81],"set":[83],"perturbations":[85],"LLM":[87],"inputs,":[88],"outputs":[90],"each":[92],"perturbation,":[93],"incorporating":[95],"an":[96],"aggregation":[97,119],"module":[98],"that":[99,144],"generalizes":[100],"approach":[104],"tasks.Through":[108],"extensive":[109],"experiments":[110],"various":[112],"datasets,":[113],"investigated":[115],"different":[116],"techniques.Our":[120],"findings":[121,142],"show":[122],"substantial":[124],"improvement":[125],"model":[127],"calibration,":[129],"reduction":[132],"Expected":[134],"Calibration":[135],"Error":[136],"(ECE)":[137],"50%":[139],"average.Our":[141],"suggest":[143],"our":[145],"proposed":[146],"offers":[149],"promising":[150],"steps":[151],"toward":[152],"enhancing":[153],"reliability":[155],"trustworthiness":[157],"LLMs":[159],"1":[160],".":[161]},"counts_by_year":[{"year":2026,"cited_by_count":5},{"year":2025,"cited_by_count":5}],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2025-10-10T00:00:00"}