{"id":"https://openalex.org/W7155217973","doi":"https://doi.org/10.48550/arxiv.2604.18811","title":"Rethinking Dataset Distillation: Hard Truths about Soft Labels","display_name":"Rethinking Dataset Distillation: Hard Truths about Soft Labels","publication_year":2026,"publication_date":"2026-04-20","ids":{"openalex":"https://openalex.org/W7155217973","doi":"https://doi.org/10.48550/arxiv.2604.18811"},"language":"en","primary_location":{"id":"pmh:doi:10.13016/m2ja89-femp","is_oa":true,"landing_page_url":"http://hdl.handle.net/11603/42941","pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Text"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"http://hdl.handle.net/11603/42941","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5019824733","display_name":"Priyam Dey","orcid":"https://orcid.org/0000-0001-5807-1379"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Dey, Priyam","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134296258","display_name":"Aditya Sahdev","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Sahdev, Aditya","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134259047","display_name":"Sunny Bhati","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Bhati, Sunny","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5024899038","display_name":"Konda Reddy Mopuri","orcid":"https://orcid.org/0000-0001-8894-7212"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Mopuri, Konda Reddy","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5134330246","display_name":"R. Venkatesh Babu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Babu, R. Venkatesh","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":0.2401999980211258,"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":0.2401999980211258,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.22779999673366547,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/T10689","display_name":"Remote-Sensing Image Classification","score":0.09889999777078629,"subfield":{"id":"https://openalex.org/subfields/2214","display_name":"Media Technology"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/scalability","display_name":"Scalability","score":0.6065999865531921},{"id":"https://openalex.org/keywords/metric","display_name":"Metric (unit)","score":0.5734000205993652},{"id":"https://openalex.org/keywords/pruning","display_name":"Pruning","score":0.536300003528595},{"id":"https://openalex.org/keywords/contrast","display_name":"Contrast (vision)","score":0.4629000127315521},{"id":"https://openalex.org/keywords/quality","display_name":"Quality (philosophy)","score":0.4431000053882599},{"id":"https://openalex.org/keywords/sample","display_name":"Sample (material)","score":0.3986000120639801},{"id":"https://openalex.org/keywords/baseline","display_name":"Baseline (sea)","score":0.38429999351501465}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6814000010490417},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.6065999865531921},{"id":"https://openalex.org/C176217482","wikidata":"https://www.wikidata.org/wiki/Q860554","display_name":"Metric (unit)","level":2,"score":0.5734000205993652},{"id":"https://openalex.org/C108010975","wikidata":"https://www.wikidata.org/wiki/Q500094","display_name":"Pruning","level":2,"score":0.536300003528595},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5257999897003174},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5242999792098999},{"id":"https://openalex.org/C2776502983","wikidata":"https://www.wikidata.org/wiki/Q690182","display_name":"Contrast (vision)","level":2,"score":0.4629000127315521},{"id":"https://openalex.org/C2779530757","wikidata":"https://www.wikidata.org/wiki/Q1207505","display_name":"Quality (philosophy)","level":2,"score":0.4431000053882599},{"id":"https://openalex.org/C198531522","wikidata":"https://www.wikidata.org/wiki/Q485146","display_name":"Sample (material)","level":2,"score":0.3986000120639801},{"id":"https://openalex.org/C12725497","wikidata":"https://www.wikidata.org/wiki/Q810247","display_name":"Baseline (sea)","level":2,"score":0.38429999351501465},{"id":"https://openalex.org/C169258074","wikidata":"https://www.wikidata.org/wiki/Q245748","display_name":"Random forest","level":2,"score":0.3441999852657318},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.34389999508857727},{"id":"https://openalex.org/C115051666","wikidata":"https://www.wikidata.org/wiki/Q6522493","display_name":"Ranging","level":2,"score":0.32179999351501465},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.29670000076293945},{"id":"https://openalex.org/C24756922","wikidata":"https://www.wikidata.org/wiki/Q1757694","display_name":"Data quality","level":3,"score":0.2928999960422516},{"id":"https://openalex.org/C129848803","wikidata":"https://www.wikidata.org/wiki/Q2564360","display_name":"Sample size determination","level":2,"score":0.29030001163482666},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.28949999809265137}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:doi:10.13016/m2ja89-femp","is_oa":true,"landing_page_url":"http://hdl.handle.net/11603/42941","pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Text"},{"id":"doi:10.48550/arxiv.2604.18811","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.18811","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":"pmh:doi:10.13016/m2ja89-femp","is_oa":true,"landing_page_url":"http://hdl.handle.net/11603/42941","pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Text"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Despite":[0],"the":[1,28,42,55,71,108,118,128,148,160,181],"perceived":[2],"success":[3],"of":[4,30,73,132,151,172,224],"large-scale":[5,174],"dataset":[6],"distillation":[7],"(DD)":[8],"methods,":[9],"recent":[10],"evidence":[11],"finds":[12],"that":[13,101,185,220],"simple":[14],"random":[15,52,109,190],"image":[16],"baselines":[17,191],"perform":[18,64],"on-par":[19],"with":[20,41],"state-of-theart":[21],"DD":[22,178,240,244,260],"methods":[23,179,201,245],"like":[24],"SRe2L":[25],"due":[26,202],"to":[27,69,89,105,127,203,235,266],"use":[29,233],"soft":[31,83,91,153],"labels":[32,84,92,96,154],"during":[33],"downstream":[34],"model":[35,156],"training.":[36],"This":[37,142],"is":[38],"in":[39,44,54,111,180],"contrast":[40],"findings":[43,254],"coreset":[45,200],"literature,":[46],"where":[47,158],"high-quality":[48,102],"coresets":[49,103,272],"consistently":[50],"outperform":[51,107],"subsets":[53],"hardlabel":[56],"(HL)":[57],"setting.":[58],"To":[59],"understand":[60],"this":[61],"discrepancy,":[62],"we":[63,213],"a":[65,138,216,228,238],"detailed":[66],"scalability":[67],"analysis":[68,99],"examine":[70],"role":[72],"data":[74],"quality":[75,164],"under":[76],"different":[77],"label":[78],"regimes,":[79],"ranging":[80],"from":[81],"abundant":[82],"(termed":[85],"as":[86],"SL+KD":[87,115,119],"regime)":[88],"fixed":[90],"(SL)":[93],"and":[94,114,175,232,262,273],"hard":[95],"(HL).":[97],"Our":[98],"reveals":[100,184],"fail":[104],"convincingly":[106],"baseline":[110],"both":[112,271],"SL":[113],"regimes.":[116],"In":[117],"setting,":[120,162],"performance":[121,143],"further":[122],"approaches":[123],"nearoptimal":[124],"levels":[125],"relative":[126],"full":[129],"dataset,":[130],"regardless":[131],"subset":[133,163],"size":[134],"or":[135],"quality,":[136],"for":[137,155,227,270],"given":[139,229],"compute":[140,230],"budget.":[141],"saturation":[144],"calls":[145],"into":[146,258],"question":[147],"widespread":[149],"practice":[150],"using":[152],"evaluation,":[157],"unlike":[159],"HL":[161,182],"has":[165],"negligible":[166],"influence.":[167],"A":[168],"subsequent":[169],"systematic":[170],"evaluation":[171],"five":[173],"four":[176],"small-scale":[177],"setting":[183],"only":[186],"RDED":[187],"reliably":[188],"outperforms":[189],"on":[192,206,211,246],"ImageNet-1K,":[193],"but":[194],"can":[195],"still":[196],"lag":[197],"behind":[198],"strong":[199],"its":[204],"over-reliance":[205],"easy":[207],"sample":[208],"patches.":[209],"Based":[210],"this,":[212],"introduce":[214],"CAD-Prune,":[215],"compute-aware":[217],"pruning":[218],"metric":[219],"efficiently":[221],"identifies":[222],"samples":[223],"optimal":[225],"difficulty":[226],"budget,":[231],"it":[234],"develop":[236],"CA2D,":[237],"compute-aligned":[239],"method,":[241],"outperforming":[242],"current":[243,259],"ImageNet-1K":[247],"at":[248],"various":[249],"IPC":[250],"settings.":[251],"Together,":[252],"our":[253],"uncover":[255],"many":[256],"insights":[257],"research":[261],"establish":[263],"useful":[264],"tools":[265],"advance":[267],"dataefficient":[268],"learning":[269],"DD.":[274]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2026-04-23T00:00:00"}