{"id":"https://openalex.org/W4412888497","doi":"https://doi.org/10.18653/v1/2025.findings-acl.515","title":"MIG: Automatic Data Selection for Instruction Tuning by Maximizing Information Gain in Semantic Space","display_name":"MIG: Automatic Data Selection for Instruction Tuning by Maximizing Information Gain in Semantic Space","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4412888497","doi":"https://doi.org/10.18653/v1/2025.findings-acl.515"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2025.findings-acl.515","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-acl.515","pdf_url":"https://aclanthology.org/2025.findings-acl.515.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: ACL 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-acl.515.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5100769177","display_name":"Yi-Cheng Chen","orcid":"https://orcid.org/0000-0003-1876-9320"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yicheng Chen","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101546250","display_name":"Yining Li","orcid":"https://orcid.org/0000-0003-4761-2293"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yining Li","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5063008981","display_name":"Kai Hu","orcid":"https://orcid.org/0000-0002-2255-6953"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Kai Hu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5119181431","display_name":"Ma Zerun","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ma Zerun","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5119181432","display_name":"HaochenYe HaochenYe","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"HaochenYe HaochenYe","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5016780403","display_name":"Kai Chen","orcid":"https://orcid.org/0009-0008-6837-3774"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Kai Chen","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":3.8492,"has_fulltext":true,"cited_by_count":2,"citation_normalized_percentile":{"value":0.93256299,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":97,"max":98},"biblio":{"volume":null,"issue":null,"first_page":"9902","last_page":"9915"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T14025","display_name":"Educational Technology and Assessment","score":0.7610999941825867,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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/T14025","display_name":"Educational Technology and Assessment","score":0.7610999941825867,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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/T11902","display_name":"Intelligent Tutoring Systems and Adaptive Learning","score":0.7161999940872192,"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":0.8029568791389465},{"id":"https://openalex.org/keywords/selection","display_name":"Selection (genetic algorithm)","score":0.7280800342559814},{"id":"https://openalex.org/keywords/space","display_name":"Space (punctuation)","score":0.5927929878234863},{"id":"https://openalex.org/keywords/information-gain","display_name":"Information gain","score":0.4872936010360718},{"id":"https://openalex.org/keywords/information-retrieval","display_name":"Information retrieval","score":0.41334348917007446},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.32455018162727356},{"id":"https://openalex.org/keywords/operating-system","display_name":"Operating system","score":0.06782633066177368}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8029568791389465},{"id":"https://openalex.org/C81917197","wikidata":"https://www.wikidata.org/wiki/Q628760","display_name":"Selection (genetic algorithm)","level":2,"score":0.7280800342559814},{"id":"https://openalex.org/C2778572836","wikidata":"https://www.wikidata.org/wiki/Q380933","display_name":"Space (punctuation)","level":2,"score":0.5927929878234863},{"id":"https://openalex.org/C2983203078","wikidata":"https://www.wikidata.org/wiki/Q255166","display_name":"Information gain","level":2,"score":0.4872936010360718},{"id":"https://openalex.org/C23123220","wikidata":"https://www.wikidata.org/wiki/Q816826","display_name":"Information retrieval","level":1,"score":0.41334348917007446},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.32455018162727356},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.06782633066177368}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2025.findings-acl.515","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-acl.515","pdf_url":"https://aclanthology.org/2025.findings-acl.515.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: ACL 2025","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2025.findings-acl.515","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-acl.515","pdf_url":"https://aclanthology.org/2025.findings-acl.515.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: ACL 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/W4412888497.pdf","grobid_xml":"https://content.openalex.org/works/W4412888497.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W2359871929","https://openalex.org/W4205762803","https://openalex.org/W2364289883","https://openalex.org/W2535856026","https://openalex.org/W2265065644","https://openalex.org/W2134699697","https://openalex.org/W3017188156","https://openalex.org/W2322875716","https://openalex.org/W2383516975","https://openalex.org/W2312656779"],"abstract_inverted_index":{"Data":[0],"quality":[1,40],"and":[2,27,41,114,154,193],"diversity":[3,116],"are":[4],"key":[5],"to":[6,23,45,60,77,142,177],"the":[7,13,55,72,80,86,99,106,119,124,144,164,178,184],"construction":[8],"of":[9,16,34,50,54,82,102,121,189],"effective":[10],"instruction-tuning":[11,18],"datasets.With":[12],"increasing":[14],"availability":[15],"open-source":[17],"datasets,":[19],"it":[20],"is":[21],"advantageous":[22],"automatically":[24],"select":[25],"high-quality":[26],"diverse":[28],"subsets":[29],"from":[30],"a":[31,51,94,111,128],"vast":[32],"amount":[33],"data.Existing":[35],"methods":[36],"typically":[37],"prioritize":[38],"instance":[39],"use":[42],"heuristic":[43,63],"rules":[44,64],"maintain":[46],"diversity.However,":[47],"this":[48,90],"absence":[49],"comprehensive":[52],"view":[53],"entire":[56],"collection":[57],"often":[58],"leads":[59],"suboptimal":[61],"results.Moreover,":[62],"generally":[65],"focus":[66],"on":[67,118,126,151,183,191,195],"distance":[68],"or":[69],"clustering":[70],"within":[71,123],"embedding":[73],"space,":[74],"which":[75],"fails":[76],"accurately":[78],"capture":[79],"intent":[81],"complex":[83],"instructions":[84],"in":[85,148],"semantic":[87,107,149],"space.To":[88],"bridge":[89],"gap,":[91],"we":[92,130],"propose":[93],"unified":[95],"method":[96,104,136],"for":[97],"quantifying":[98],"information":[100,122],"content":[101],"datasets.This":[103],"models":[105,156],"space":[108],"by":[109,172],"constructing":[110],"label":[112],"graph":[113],"quantifies":[115],"based":[117],"distribution":[120],"graph.Based":[125],"such":[127],"measurement,":[129],"further":[131],"introduce":[132],"an":[133],"efficient":[134],"sampling":[135],"that":[137,158],"selects":[138],"data":[139,170],"samples":[140],"iteratively":[141],"Maximize":[143],"Information":[145],"Gain":[146],"(MIG)":[147],"space.Experiments":[150],"various":[152],"datasets":[153],"base":[155],"demonstrate":[157],"MIG":[159,173],"consistently":[160],"outperforms":[161],"state-of-the-art":[162],"methods.Notably,":[163],"model":[165,181],"fine-tuned":[166],"with":[167,187],"5%":[168],"Tulu3":[169],"sampled":[171],"achieves":[174],"comparable":[175],"performance":[176],"official":[179],"SFT":[180],"trained":[182],"full":[185],"dataset,":[186],"improvements":[188],"+5.73%":[190],"AlpacaEval":[192],"+6.89%":[194],"Wildbench.":[196]},"counts_by_year":[{"year":2026,"cited_by_count":2}],"updated_date":"2026-07-29T14:22:42.915294","created_date":"2025-10-10T00:00:00"}