{"id":"https://openalex.org/W4416034960","doi":"https://doi.org/10.18653/v1/2025.findings-emnlp.226","title":"Benchmarking Uncertainty Metrics for LLM Target-Aware Search","display_name":"Benchmarking Uncertainty Metrics for LLM Target-Aware Search","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4416034960","doi":"https://doi.org/10.18653/v1/2025.findings-emnlp.226"},"language":null,"primary_location":{"id":"doi:10.18653/v1/2025.findings-emnlp.226","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-emnlp.226","pdf_url":"https://aclanthology.org/2025.findings-emnlp.226.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":"Findings of the Association for Computational Linguistics: EMNLP 2025","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2025.findings-emnlp.226.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5109683004","display_name":"Pei-Fu Guo","orcid":null},"institutions":[{"id":"https://openalex.org/I16733864","display_name":"National Taiwan University","ror":"https://ror.org/05bqach95","country_code":"TW","type":"education","lineage":["https://openalex.org/I16733864"]}],"countries":["TW"],"is_corresponding":false,"raw_author_name":"Pei-Fu Guo","raw_affiliation_strings":["National Taiwan University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"National Taiwan University","institution_ids":["https://openalex.org/I16733864"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5114386402","display_name":"Yun-Da Tsai","orcid":null},"institutions":[{"id":"https://openalex.org/I16733864","display_name":"National Taiwan University","ror":"https://ror.org/05bqach95","country_code":"TW","type":"education","lineage":["https://openalex.org/I16733864"]}],"countries":["TW"],"is_corresponding":false,"raw_author_name":"Yun-Da Tsai","raw_affiliation_strings":["National Taiwan University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"National Taiwan University","institution_ids":["https://openalex.org/I16733864"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5087480257","display_name":"Shou-De Lin","orcid":"https://orcid.org/0000-0001-9970-1250"},"institutions":[{"id":"https://openalex.org/I16733864","display_name":"National Taiwan University","ror":"https://ror.org/05bqach95","country_code":"TW","type":"education","lineage":["https://openalex.org/I16733864"]}],"countries":["TW"],"is_corresponding":false,"raw_author_name":"Shou-De Lin","raw_affiliation_strings":["National Taiwan University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"National Taiwan University","institution_ids":["https://openalex.org/I16733864"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I16733864"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.39633014,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"4230","last_page":"4238"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10317","display_name":"Advanced Database Systems and Queries","score":0.15569999814033508,"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"}},"topics":[{"id":"https://openalex.org/T10317","display_name":"Advanced Database Systems and Queries","score":0.15569999814033508,"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"}},{"id":"https://openalex.org/T11719","display_name":"Data Quality and Management","score":0.15230000019073486,"subfield":{"id":"https://openalex.org/subfields/1803","display_name":"Management Science and Operations Research"},"field":{"id":"https://openalex.org/fields/18","display_name":"Decision Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}},{"id":"https://openalex.org/T11106","display_name":"Data Management and Algorithms","score":0.14480000734329224,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/benchmarking","display_name":"Benchmarking","score":0.5613999962806702},{"id":"https://openalex.org/keywords/measure","display_name":"Measure (data warehouse)","score":0.3668999969959259},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.2939000129699707},{"id":"https://openalex.org/keywords/benchmark","display_name":"Benchmark (surveying)","score":0.2759999930858612},{"id":"https://openalex.org/keywords/probabilistic-logic","display_name":"Probabilistic logic","score":0.26339998841285706}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5834000110626221},{"id":"https://openalex.org/C86251818","wikidata":"https://www.wikidata.org/wiki/Q816754","display_name":"Benchmarking","level":2,"score":0.5613999962806702},{"id":"https://openalex.org/C2780009758","wikidata":"https://www.wikidata.org/wiki/Q6804172","display_name":"Measure (data warehouse)","level":2,"score":0.3668999969959259},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.3481000065803528},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.2939000129699707},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.2759999930858612},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.2721000015735626},{"id":"https://openalex.org/C49937458","wikidata":"https://www.wikidata.org/wiki/Q2599292","display_name":"Probabilistic logic","level":2,"score":0.26339998841285706},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.2467000037431717},{"id":"https://openalex.org/C176217482","wikidata":"https://www.wikidata.org/wiki/Q860554","display_name":"Metric (unit)","level":2,"score":0.24529999494552612}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2025.findings-emnlp.226","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-emnlp.226","pdf_url":"https://aclanthology.org/2025.findings-emnlp.226.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":"Findings of the Association for Computational Linguistics: EMNLP 2025","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2025.findings-emnlp.226","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-emnlp.226","pdf_url":"https://aclanthology.org/2025.findings-emnlp.226.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":"Findings of the Association for Computational Linguistics: EMNLP 2025","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4416034960.pdf","grobid_xml":"https://content.openalex.org/works/W4416034960.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"LLM":[0,15,52,136],"search":[1,24,45,66,115],"methods,":[2,30],"such":[3],"as":[4],"Chain":[5],"of":[6,11,60,106],"Thought":[7,12],"(CoT)":[8],"and":[9,28,85],"Tree":[10],"(ToT),":[13],"enhance":[14],"reasoning":[16,20],"by":[17],"exploring":[18],"multiple":[19],"paths.When":[21],"combined":[22],"with":[23,43,102],"algorithms":[25],"like":[26],"MCTS":[27],"Bandit":[29],"their":[31,111],"effectiveness":[32,112],"relies":[33],"heavily":[34],"on":[35],"uncertainty":[36,53,61,75,80,108,126],"estimation":[37],"to":[38,63,131],"prioritize":[39],"paths":[40],"that":[41,97,128],"align":[42,101],"specific":[44],"objectives.However,":[46],"it":[47],"remains":[48],"unclear":[49],"whether":[50],"existing":[51],"metrics":[54,99],"adequately":[55],"capture":[56],"the":[57,121],"diverse":[58],"types":[59],"required":[62],"guide":[64],"different":[65,90],"objectives.In":[67],"this":[68],"work,":[69],"we":[70],"introduce":[71],"a":[72,104],"framework":[73],"for":[74,113,123],"benchmarking,":[76],"identifying":[77],"four":[78],"distinct":[79],"types:":[81],"Answer,":[82],"Correctness,":[83],"Aleatoric,":[84],"Epistemic":[86],"Uncertainty.Each":[87],"type":[88],"serves":[89],"optimization":[91,133],"goals":[92,134],"in":[93,116,135],"search.Our":[94],"experiments":[95],"demonstrate":[96],"current":[98],"often":[100],"only":[103],"subset":[105],"these":[107],"types,":[109],"limiting":[110],"objective-aligned":[114],"some":[117],"cases.These":[118],"findings":[119],"highlight":[120],"need":[122],"additional":[124],"target-aware":[125],"estimators":[127],"can":[129],"adapt":[130],"various":[132],"search.Code":[137],"is":[138],"available":[139],"at":[140],"link.":[141]},"counts_by_year":[],"updated_date":"2026-07-29T14:22:42.915294","created_date":"2025-11-08T00:00:00"}