{"id":"https://openalex.org/W7161740963","doi":"https://doi.org/10.48550/arxiv.2605.18654","title":"Pocket Foundation Models: Distilling TFMs into CPU-Ready Gradient-Boosted Trees","display_name":"Pocket Foundation Models: Distilling TFMs into CPU-Ready Gradient-Boosted Trees","publication_year":2026,"publication_date":"2026-05-18","ids":{"openalex":"https://openalex.org/W7161740963","doi":"https://doi.org/10.48550/arxiv.2605.18654"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2605.18654","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.18654","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","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.2605.18654","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5136494900","display_name":"Aditya Tanna","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tanna, Aditya","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5016199207","display_name":"Nassim Bouarour","orcid":"https://orcid.org/0000-0003-1202-351X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Bouarour, Nassim","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5099022862","display_name":"Mohamed Bouadi","orcid":"https://orcid.org/0009-0001-3905-0588"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Bouadi, Mohamed","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136471029","display_name":"Vinay Kumar Sankarapu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Sankarapu, Vinay kumar","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5012713026","display_name":"Pratinav Seth","orcid":"https://orcid.org/0009-0001-4525-4464"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Seth, Pratinav","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/T12535","display_name":"Machine Learning and Data Classification","score":0.2224999964237213,"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":0.2224999964237213,"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/T11652","display_name":"Imbalanced Data Classification Techniques","score":0.13500000536441803,"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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.1340000033378601,"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/pipeline","display_name":"Pipeline (software)","score":0.763700008392334},{"id":"https://openalex.org/keywords/tree","display_name":"Tree (set theory)","score":0.5254999995231628},{"id":"https://openalex.org/keywords/speedup","display_name":"Speedup","score":0.5180000066757202},{"id":"https://openalex.org/keywords/baseline","display_name":"Baseline (sea)","score":0.4986000061035156},{"id":"https://openalex.org/keywords/rank","display_name":"Rank (graph theory)","score":0.4984999895095825},{"id":"https://openalex.org/keywords/obstacle","display_name":"Obstacle","score":0.4611999988555908},{"id":"https://openalex.org/keywords/foundation","display_name":"Foundation (evidence)","score":0.4415000081062317},{"id":"https://openalex.org/keywords/enhanced-data-rates-for-gsm-evolution","display_name":"Enhanced Data Rates for GSM Evolution","score":0.427700012922287}],"concepts":[{"id":"https://openalex.org/C43521106","wikidata":"https://www.wikidata.org/wiki/Q2165493","display_name":"Pipeline (software)","level":2,"score":0.763700008392334},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6026999950408936},{"id":"https://openalex.org/C113174947","wikidata":"https://www.wikidata.org/wiki/Q2859736","display_name":"Tree (set theory)","level":2,"score":0.5254999995231628},{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.5180000066757202},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5138999819755554},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5020999908447266},{"id":"https://openalex.org/C12725497","wikidata":"https://www.wikidata.org/wiki/Q810247","display_name":"Baseline (sea)","level":2,"score":0.4986000061035156},{"id":"https://openalex.org/C164226766","wikidata":"https://www.wikidata.org/wiki/Q7293202","display_name":"Rank (graph theory)","level":2,"score":0.4984999895095825},{"id":"https://openalex.org/C2776650193","wikidata":"https://www.wikidata.org/wiki/Q264661","display_name":"Obstacle","level":2,"score":0.4611999988555908},{"id":"https://openalex.org/C2780966255","wikidata":"https://www.wikidata.org/wiki/Q5474306","display_name":"Foundation (evidence)","level":2,"score":0.4415000081062317},{"id":"https://openalex.org/C162307627","wikidata":"https://www.wikidata.org/wiki/Q204833","display_name":"Enhanced Data Rates for GSM Evolution","level":2,"score":0.427700012922287},{"id":"https://openalex.org/C108010975","wikidata":"https://www.wikidata.org/wiki/Q500094","display_name":"Pruning","level":2,"score":0.3831000030040741},{"id":"https://openalex.org/C84525736","wikidata":"https://www.wikidata.org/wiki/Q831366","display_name":"Decision tree","level":2,"score":0.3601999878883362},{"id":"https://openalex.org/C204030448","wikidata":"https://www.wikidata.org/wiki/Q101017","display_name":"Distillation","level":2,"score":0.31360000371932983},{"id":"https://openalex.org/C86037889","wikidata":"https://www.wikidata.org/wiki/Q4330127","display_name":"Learning to rank","level":3,"score":0.305400013923645},{"id":"https://openalex.org/C107094494","wikidata":"https://www.wikidata.org/wiki/Q428453","display_name":"Fault tree analysis","level":2,"score":0.29679998755455017},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.2863999903202057},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.28209999203681946},{"id":"https://openalex.org/C47177190","wikidata":"https://www.wikidata.org/wiki/Q207137","display_name":"Curriculum","level":2,"score":0.2621000111103058},{"id":"https://openalex.org/C198352243","wikidata":"https://www.wikidata.org/wiki/Q37105","display_name":"Line (geometry)","level":2,"score":0.25949999690055847},{"id":"https://openalex.org/C2781002164","wikidata":"https://www.wikidata.org/wiki/Q6822311","display_name":"Meta learning (computer science)","level":3,"score":0.25920000672340393},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.257999986410141},{"id":"https://openalex.org/C175309249","wikidata":"https://www.wikidata.org/wiki/Q725864","display_name":"Pipeline transport","level":2,"score":0.2515000104904175}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2605.18654","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.18654","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2605.18654","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.18654","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"sustainable_development_goals":[{"id":"https://metadata.un.org/sdg/16","display_name":"Peace, Justice and strong institutions","score":0.7031830549240112}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"A":[0],"fraud":[1],"scorer":[2],"needs":[3],"to":[4,46,65,73,112,142],"answer":[5],"in":[6],"under":[7],"2":[8],"ms.":[9],"The":[10,41,194],"best":[11],"tabular":[12],"foundation":[13],"models":[14],"(TFMs)":[15],"take":[16],"151-1,275":[17],"ms":[18,107],"on":[19,39,108,147,178],"GPU.":[20],"We":[21],"close":[22],"this":[23],"gap":[24],"by":[25],"distilling":[26,93],"the":[27,61,182,202],"TFM":[28],"offline":[29],"into":[30,95],"an":[31],"XGBoost":[32,96],"or":[33],"CatBoost":[34,126,155],"student":[35,143],"that":[36],"runs":[37],"natively":[38],"CPU.":[40],"central":[42],"obstacle":[43],"is":[44,197],"specific":[45],"in-context":[47],"learning":[48],"(ICL)":[49],"teachers:":[50],"they":[51],"leak":[52],"labels":[53],"when":[54],"scoring":[55],"their":[56],"own":[57],"training":[58],"set,":[59],"so":[60],"soft":[62],"targets":[63],"collapse":[64],"near-one-hot":[66],"vectors":[67],"with":[68,118],"no":[69],"inter-class":[70],"structure":[71],"left":[72],"distill.":[74],"Stratified":[75],"out-of-fold":[76],"(OOF)":[77],"teacher":[78,103,138,183],"labeling":[79],"prevents":[80],"this.":[81],"Across":[82],"153":[83],"classification":[84],"datasets":[85],"drawn":[86],"from":[87],"TALENT,":[88],"OpenML-CC18,":[89],"TabZilla,":[90],"and":[91,177],"TabArena,":[92],"TabICLv2":[94],"gives":[97],"0.882":[98],"macro-mean":[99],"AUC":[100],"(96.5%":[101],"of":[102,201],"AUC)":[104],"at":[105],"1.9":[106],"CPU,":[109],"a":[110,119,124],"38x":[111],"860x":[113],"speedup":[114],"across":[115],"teacher-student":[116],"pairs":[117],"statistically":[120],"significant":[121],"edge":[122],"over":[123,154],"tuned":[125],"baseline":[127],"(Wilcoxon":[128],"p":[129,166],"=":[130,167],"0.0008;":[131],"51%":[132],"win":[133],"rate).":[134],"Four":[135],"further":[136],"findings:":[137],"rank":[139],"transfers":[140],"exactly":[141],"rank;":[144],"gains":[145],"concentrate":[146],"low-dimensional":[148],"data":[149],"(<":[150],"21":[151],"features:":[152,158],"+0.011":[153],"vs.":[156],">21":[157],"+0.001);":[159],"multi-teacher":[160],"averaging":[161],"helps":[162],"MLP":[163],"students":[164],"(+0.006,":[165],"0.003)":[168],"but":[169],"adds":[170],"less":[171],"than":[172,192],"0.001":[173],"for":[174],"tree":[175],"students;":[176],"high-dimensional":[179],"tasks":[180],"where":[181],"itself":[184],"trails":[185],"CatBoost,":[186],"distillation":[187],"makes":[188],"things":[189],"worse":[190],"rather":[191],"better.":[193],"full":[195],"pipeline":[196],"open-sourced":[198],"as":[199],"part":[200],"TabTune":[203],"library.":[204]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2026-05-20T00:00:00"}