{"id":"https://openalex.org/W7164029780","doi":"https://doi.org/10.48550/arxiv.2606.08167","title":"Explaining Data Mixing Scaling Laws","display_name":"Explaining Data Mixing Scaling Laws","publication_year":2026,"publication_date":"2026-06-06","ids":{"openalex":"https://openalex.org/W7164029780","doi":"https://doi.org/10.48550/arxiv.2606.08167"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2606.08167","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.08167","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"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":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2606.08167","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5138234195","display_name":"Rui Dai","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Dai, Rui","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5067116235","display_name":"Shuran Zheng","orcid":"https://orcid.org/0000-0003-2575-8283"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zheng, Shuran","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":null,"has_fulltext":false,"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.2962999939918518,"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.2962999939918518,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.12929999828338623,"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":0.10159999877214432,"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/scaling","display_name":"Scaling","score":0.6399999856948853},{"id":"https://openalex.org/keywords/mixing","display_name":"Mixing (physics)","score":0.4900999963283539},{"id":"https://openalex.org/keywords/domain","display_name":"Domain (mathematical analysis)","score":0.4731000065803528},{"id":"https://openalex.org/keywords/scaling-law","display_name":"Scaling law","score":0.45899999141693115},{"id":"https://openalex.org/keywords/empirical-research","display_name":"Empirical research","score":0.451200008392334},{"id":"https://openalex.org/keywords/empirical-modelling","display_name":"Empirical modelling","score":0.4147000014781952},{"id":"https://openalex.org/keywords/code","display_name":"Code (set theory)","score":0.3962000012397766}],"concepts":[{"id":"https://openalex.org/C99844830","wikidata":"https://www.wikidata.org/wiki/Q102441924","display_name":"Scaling","level":2,"score":0.6399999856948853},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5076000094413757},{"id":"https://openalex.org/C138777275","wikidata":"https://www.wikidata.org/wiki/Q6884054","display_name":"Mixing (physics)","level":2,"score":0.4900999963283539},{"id":"https://openalex.org/C36503486","wikidata":"https://www.wikidata.org/wiki/Q11235244","display_name":"Domain (mathematical analysis)","level":2,"score":0.4731000065803528},{"id":"https://openalex.org/C2988430800","wikidata":"https://www.wikidata.org/wiki/Q428971","display_name":"Scaling law","level":3,"score":0.45899999141693115},{"id":"https://openalex.org/C120936955","wikidata":"https://www.wikidata.org/wiki/Q2155640","display_name":"Empirical research","level":2,"score":0.451200008392334},{"id":"https://openalex.org/C133199616","wikidata":"https://www.wikidata.org/wiki/Q25386885","display_name":"Empirical modelling","level":2,"score":0.4147000014781952},{"id":"https://openalex.org/C121864883","wikidata":"https://www.wikidata.org/wiki/Q677916","display_name":"Statistical physics","level":1,"score":0.3978999853134155},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.3962000012397766},{"id":"https://openalex.org/C149782125","wikidata":"https://www.wikidata.org/wiki/Q160039","display_name":"Econometrics","level":1,"score":0.3555000126361847},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.3522000014781952},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.3424000144004822},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.30720001459121704},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.2897000014781952},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.28700000047683716},{"id":"https://openalex.org/C103824480","wikidata":"https://www.wikidata.org/wiki/Q185889","display_name":"Time domain","level":2,"score":0.28299999237060547},{"id":"https://openalex.org/C28826006","wikidata":"https://www.wikidata.org/wiki/Q33521","display_name":"Applied mathematics","level":1,"score":0.28029999136924744},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.2800000011920929},{"id":"https://openalex.org/C55037315","wikidata":"https://www.wikidata.org/wiki/Q5421151","display_name":"Experimental data","level":2,"score":0.27230000495910645},{"id":"https://openalex.org/C2778755073","wikidata":"https://www.wikidata.org/wiki/Q10858537","display_name":"Scale (ratio)","level":2,"score":0.2709999978542328},{"id":"https://openalex.org/C45493050","wikidata":"https://www.wikidata.org/wiki/Q7884934","display_name":"Unified Model","level":2,"score":0.2639000117778778},{"id":"https://openalex.org/C61797465","wikidata":"https://www.wikidata.org/wiki/Q1188986","display_name":"Term (time)","level":2,"score":0.26179999113082886},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.2612000107765198}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2606.08167","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.08167","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"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":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2606.08167","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.08167","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"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":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"sustainable_development_goals":[{"display_name":"Quality Education","score":0.6759635210037231,"id":"https://metadata.un.org/sdg/4"}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Recent":[0],"research":[1],"has":[2],"established":[3],"empirical":[4,183],"scaling":[5,52],"laws":[6,53],"to":[7,34,58,118,181],"predict":[8],"model":[9,21,102,150,172],"performance":[10],"on":[11,63,70,75,91,166],"multi-domain":[12,60],"data":[13,40,93],"mixtures.":[14,146],"However,":[15],"a":[16,31,137],"theoretical":[17,45],"understanding":[18],"of":[19,39,88,100],"these":[20,174],"loss":[22,134],"behaviors":[23],"remains":[24],"absent.":[25],"In":[26,169],"this":[27],"work,":[28],"we":[29,78],"propose":[30],"unified":[32],"framework":[33,127],"explain":[35],"the":[36,59,64,85,98,133],"underlying":[37],"mechanics":[38],"mixing.":[41],"Our":[42,185],"approach":[43],"extends":[44],"perspectives":[46],"originally":[47],"developed":[48],"for":[49,159],"standard":[50],"neural":[51],"(e.g.,":[54],"Kaplan":[55],"and":[56,108,142],"Chinchilla)":[57],"setting.":[61],"Based":[62],"distributional":[65],"assumption":[66],"that":[67,83,125],"domains":[68,117],"overlap":[69],"fundamental":[71],"skills":[72],"while":[73],"diverging":[74],"specialized":[76],"skills,":[77],"identify":[79],"two":[80],"key":[81],"factors":[82],"govern":[84],"domain":[86,105],"losses":[87,106],"models":[89],"trained":[90],"different":[92],"mixtures:":[94],"\\textit{Capacity":[95],"Competition},":[96],"where":[97,111],"allocation":[99],"finite":[101],"capacity":[103],"couples":[104],"globally,":[107],"\\textit{Noise":[109],"Reduction},":[110],"optimal":[112],"weights":[113],"shift":[114],"toward":[115],"harder-to-learn":[116],"minimize":[119],"overall":[120],"noise.":[121],"Empirical":[122],"evaluations":[123],"show":[124],"our":[126,149,171],"outperforms":[128],"existing":[129],"baselines":[130],"by":[131],"fitting":[132],"landscape":[135],"with":[136],"lower":[138],"Mean":[139],"Relative":[140],"Error":[141],"identifying":[143],"higher-performing":[144],"training":[145],"Most":[147],"importantly,":[148],"successfully":[151],"extrapolates":[152],"across":[153],"scales,":[154],"predicting":[155],"highly":[156],"effective":[157],"mixtures":[158],"large,":[160],"unseen":[161],"scales":[162],"using":[163,176],"parameters":[164,179],"fitted":[165],"smaller":[167],"ones.":[168],"addition,":[170],"achieves":[173],"results":[175],"significantly":[177],"fewer":[178],"compared":[180],"previous":[182],"laws.":[184],"code":[186],"is":[187],"available":[188],"at":[189],"https://github.com/meiqwq/Explaining-Data-Mixing-Scaling-Laws.":[190]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2026-06-10T00:00:00"}