{"id":"https://openalex.org/W4407093498","doi":"https://doi.org/10.48550/arxiv.2501.18962","title":"Spend Wisely: Maximizing Post-Training Gains in Iterative Synthetic Data Bootstrapping","display_name":"Spend Wisely: Maximizing Post-Training Gains in Iterative Synthetic Data Bootstrapping","publication_year":2025,"publication_date":"2025-01-31","ids":{"openalex":"https://openalex.org/W4407093498","doi":"https://doi.org/10.48550/arxiv.2501.18962"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2501.18962","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2501.18962","pdf_url":"https://arxiv.org/pdf/2501.18962","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":null},"type":"preprint","indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2501.18962","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5108538129","display_name":"Pingchuan Yang","orcid":"https://orcid.org/0009-0005-2347-6903"},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Yang, Pu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5064491965","display_name":"Yunzhen Feng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Feng, Yunzhen","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101571870","display_name":"Ziyuan Chen","orcid":"https://orcid.org/0000-0003-2832-945X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Ziyuan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5001839348","display_name":"Yuhang Wu","orcid":"https://orcid.org/0000-0001-9860-1434"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wu, Yuhang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5033374865","display_name":"Zhuoyuan Li","orcid":"https://orcid.org/0000-0001-6105-6540"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Zhuoyuan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":5,"corresponding_author_ids":["https://openalex.org/A5108538129"],"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/T12535","display_name":"Machine Learning and Data Classification","score":0.9628999829292297,"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/T12535","display_name":"Machine Learning and Data Classification","score":0.9628999829292297,"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/T13702","display_name":"Machine Learning in Healthcare","score":0.946399986743927,"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/T10320","display_name":"Neural Networks and Applications","score":0.9462000131607056,"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/training","display_name":"Training (meteorology)","score":0.6619418263435364},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.5384533405303955},{"id":"https://openalex.org/keywords/training-set","display_name":"Training set","score":0.4263978600502014},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.315104603767395},{"id":"https://openalex.org/keywords/geography","display_name":"Geography","score":0.076610267162323}],"concepts":[{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.6619418263435364},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5384533405303955},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.4263978600502014},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.315104603767395},{"id":"https://openalex.org/C205649164","wikidata":"https://www.wikidata.org/wiki/Q1071","display_name":"Geography","level":0,"score":0.076610267162323},{"id":"https://openalex.org/C153294291","wikidata":"https://www.wikidata.org/wiki/Q25261","display_name":"Meteorology","level":1,"score":0.0}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2501.18962","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2501.18962","pdf_url":"https://arxiv.org/pdf/2501.18962","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":null},{"id":"doi:10.48550/arxiv.2501.18962","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2501.18962","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":null,"raw_source_name":null,"raw_type":"article"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2501.18962","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2501.18962","pdf_url":"https://arxiv.org/pdf/2501.18962","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":null},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W230091440","https://openalex.org/W2390279801","https://openalex.org/W2233261550","https://openalex.org/W4391913857","https://openalex.org/W2358668433","https://openalex.org/W4394050964","https://openalex.org/W2551249631"],"abstract_inverted_index":{"Modern":[0],"foundation":[1],"models":[2,105,112],"often":[3,128],"undergo":[4],"iterative":[5],"``bootstrapping''":[6],"in":[7],"their":[8],"post-training":[9],"phase:":[10],"a":[11,40,65],"model":[12,36],"generates":[13],"synthetic":[14],"data,":[15],"an":[16],"external":[17],"verifier":[18],"filters":[19],"out":[20],"low-quality":[21],"samples,":[22],"and":[23,50,106,117],"the":[24,35,45],"high-quality":[25],"subset":[26],"is":[27],"used":[28],"for":[29,48,68],"further":[30],"fine-tuning.":[31],"Over":[32],"multiple":[33],"iterations,":[34],"performance":[37],"improves,":[38],"raising":[39],"crucial":[41],"question:":[42],"How":[43],"should":[44],"total":[46],"budget":[47,70],"generation":[49],"training":[51],"be":[52],"allocated":[53],"across":[54],"iterations":[55],"to":[56,80],"maximize":[57],"final":[58],"performance?":[59],"In":[60],"this":[61],"work,":[62],"we":[63,74],"develop":[64],"theoretical":[66,96],"framework":[67],"analyzing":[69],"allocation":[71],"strategies.":[72],"Specifically,":[73],"show":[75,113],"that":[76,114],"constant":[77,123],"policies":[78,87,92,120,127],"fail":[79],"converge":[81],"with":[82,102,109,125],"high":[83],"probability,":[84],"while":[85],"increasing":[86],"--":[88,93],"particularly":[89],"exponential":[90,116,126],"growth":[91,119],"exhibit":[94],"significant":[95],"advantages.":[97],"Experiments":[98],"on":[99],"image":[100],"denoising":[101],"diffusion":[103],"probabilistic":[104],"math":[107],"reasoning":[108],"large":[110],"language":[111],"both":[115],"polynomial":[118],"consistently":[121],"outperform":[122],"policies,":[124],"providing":[129],"more":[130],"stable":[131],"performance.":[132]},"counts_by_year":[],"updated_date":"2026-06-06T09:05:17.133730","created_date":"2025-02-04T00:00:00"}