{"id":"https://openalex.org/W4416034350","doi":"https://doi.org/10.18653/v1/2025.findings-emnlp.621","title":"Efficiently Selecting Response Generation Strategies for Synthetic Data Construction by Self-Aligned Perplexity","display_name":"Efficiently Selecting Response Generation Strategies for Synthetic Data Construction by Self-Aligned Perplexity","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4416034350","doi":"https://doi.org/10.18653/v1/2025.findings-emnlp.621"},"language":null,"primary_location":{"id":"doi:10.18653/v1/2025.findings-emnlp.621","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-emnlp.621","pdf_url":"https://aclanthology.org/2025.findings-emnlp.621.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.621.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5114085085","display_name":"Xuan Ren","orcid":null},"institutions":[{"id":"https://openalex.org/I5681781","display_name":"The University of Adelaide","ror":"https://ror.org/00892tw58","country_code":"AU","type":"education","lineage":["https://openalex.org/I5681781"]}],"countries":["AU"],"is_corresponding":false,"raw_author_name":"Xuan Ren","raw_affiliation_strings":["University of Adelaide University of Adelaide University of Adelaide"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Adelaide University of Adelaide University of Adelaide","institution_ids":["https://openalex.org/I5681781"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102961064","display_name":"Qi Chen","orcid":"https://orcid.org/0009-0006-5926-7457"},"institutions":[{"id":"https://openalex.org/I5681781","display_name":"The University of Adelaide","ror":"https://ror.org/00892tw58","country_code":"AU","type":"education","lineage":["https://openalex.org/I5681781"]}],"countries":["AU"],"is_corresponding":false,"raw_author_name":"Qi Chen","raw_affiliation_strings":["University of Adelaide University of Adelaide University of Adelaide"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Adelaide University of Adelaide University of Adelaide","institution_ids":["https://openalex.org/I5681781"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5070976480","display_name":"Lingqiao Liu","orcid":"https://orcid.org/0000-0003-3584-795X"},"institutions":[{"id":"https://openalex.org/I5681781","display_name":"The University of Adelaide","ror":"https://ror.org/00892tw58","country_code":"AU","type":"education","lineage":["https://openalex.org/I5681781"]}],"countries":["AU"],"is_corresponding":false,"raw_author_name":"Lingqiao Liu","raw_affiliation_strings":["University of Adelaide University of Adelaide University of Adelaide"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Adelaide University of Adelaide University of Adelaide","institution_ids":["https://openalex.org/I5681781"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I5681781"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.29819205,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"11584","last_page":"11605"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10201","display_name":"Speech Recognition and Synthesis","score":0.08229999989271164,"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/T10201","display_name":"Speech Recognition and Synthesis","score":0.08229999989271164,"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/T12029","display_name":"DNA and Biological Computing","score":0.07249999791383743,"subfield":{"id":"https://openalex.org/subfields/1312","display_name":"Molecular Biology"},"field":{"id":"https://openalex.org/fields/13","display_name":"Biochemistry, Genetics and Molecular Biology"},"domain":{"id":"https://openalex.org/domains/1","display_name":"Life Sciences"}},{"id":"https://openalex.org/T10860","display_name":"Speech and Audio Processing","score":0.0502999983727932,"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/synthetic-data","display_name":"Synthetic data","score":0.5751000046730042},{"id":"https://openalex.org/keywords/perplexity","display_name":"Perplexity","score":0.5267999768257141},{"id":"https://openalex.org/keywords/selection","display_name":"Selection (genetic algorithm)","score":0.31189998984336853},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.2669000029563904},{"id":"https://openalex.org/keywords/real-world-data","display_name":"Real world data","score":0.2614000141620636}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6345000267028809},{"id":"https://openalex.org/C160920958","wikidata":"https://www.wikidata.org/wiki/Q7662746","display_name":"Synthetic data","level":2,"score":0.5751000046730042},{"id":"https://openalex.org/C100279451","wikidata":"https://www.wikidata.org/wiki/Q372193","display_name":"Perplexity","level":3,"score":0.5267999768257141},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4447000026702881},{"id":"https://openalex.org/C81917197","wikidata":"https://www.wikidata.org/wiki/Q628760","display_name":"Selection (genetic algorithm)","level":2,"score":0.31189998984336853},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.31119999289512634},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.2669000029563904},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.26660001277923584},{"id":"https://openalex.org/C3020493868","wikidata":"https://www.wikidata.org/wiki/Q55631277","display_name":"Real world data","level":2,"score":0.2614000141620636},{"id":"https://openalex.org/C168167062","wikidata":"https://www.wikidata.org/wiki/Q1117970","display_name":"Component (thermodynamics)","level":2,"score":0.2459000051021576}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2025.findings-emnlp.621","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-emnlp.621","pdf_url":"https://aclanthology.org/2025.findings-emnlp.621.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.621","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-emnlp.621","pdf_url":"https://aclanthology.org/2025.findings-emnlp.621.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/W4416034350.pdf","grobid_xml":"https://content.openalex.org/works/W4416034350.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Fine-tuning":[0],"large":[1,9],"language":[2],"models":[3],"(LLMs)":[4],"typically":[5],"relies":[6],"on":[7,41,90,186,202],"producing":[8],"sets":[10],"of":[11,64,105],"inputoutput":[12],"pairs.Yet":[13],"for":[14,112],"a":[15,96,102,113,158,187],"given":[16],"question,":[17],"there":[18],"can":[19,38,59,179],"be":[20,124,141],"many":[21],"valid":[22],"outputs.In":[23],"practice,":[24],"these":[25,54],"outputs":[26,56,122,165],"are":[27,57],"often":[28],"derived":[29],"by":[30,205],"distilling":[31],"knowledge":[32],"from":[33,81],"teacher":[34,44],"models,":[35],"and":[36,88,149,173],"they":[37],"vary":[39],"depending":[40],"the":[42,62,65,76,127,168,181,195,206],"specific":[43,114],"model":[45],"or":[46],"prompting":[47],"strategy":[48,185],"employed.Recent":[49],"findings":[50],"show":[51],"that":[52,100,120,138,200],"how":[53,72,162],"training":[55,87,197,201,227],"generated":[58,106,204],"significantly":[60],"affect":[61],"performance":[63],"fine-tuned":[66],"model,":[67],"raising":[68],"an":[69],"important":[70],"question:":[71],"do":[73],"we":[74,136,154,178],"pick":[75],"best":[77],"data":[78,107,203],"generation":[79,184],"method":[80,99,208],"among":[82],"numerous":[83],"possibilities?Rather":[84],"than":[85],"exhaustively":[86],"evaluating":[89],"each":[91],"candidate,":[92],"this":[93,176],"paper":[94],"proposes":[95],"scalable":[97],"approximate":[98],"assesses":[101],"small":[103,188],"subset":[104],"to":[108,126,167,193,224],"estimate":[109],"its":[110],"suitability":[111],"target":[115,128,169],"LLM.Our":[116],"central":[117],"idea":[118],"is":[119,230],"effective":[121,183],"should":[123],"familiar":[125],"LLM.While":[129],"previous":[130],"work":[131],"measures":[132],"familiarity":[133],"with":[134],"perplexity,":[135,157],"find":[137],"perplexity":[139],"might":[140],"suboptimal":[142],"in":[143,217],"characterizing":[144],"\"familiarity\"":[145],"through":[146],"empirical":[147],"analyses":[148],"practical":[150],"observations.To":[151],"address":[152],"this,":[153],"introduce":[155],"self-aligned":[156],"novel":[159],"metric":[160],"capturing":[161],"closely":[163],"candidate":[164,221],"adhere":[166],"LLM's":[170],"own":[171],"style":[172],"reasoning":[174],"patterns.In":[175],"way,":[177],"identify":[180],"most":[182],"sample,":[189],"then":[190],"apply":[191],"it":[192],"produce":[194],"complete":[196],"set.We":[198],"demonstrate":[199],"chosen":[207],"yields":[209],"significant":[210],"improvements":[211],"across":[212],"diverse":[213],"reasoning-focused":[214],"benchmarks,":[215],"particularly":[216],"cases":[218],"where":[219],"different":[220],"methods":[222],"lead":[223],"highly":[225],"divergent":[226],"outcomes.Our":[228],"implementation":[229],"publicly":[231],"available":[232],"at":[233]},"counts_by_year":[],"updated_date":"2026-07-29T14:22:42.915294","created_date":"2025-11-08T00:00:00"}