{"id":"https://openalex.org/W7160620685","doi":"https://doi.org/10.48550/arxiv.2605.06314","title":"When Does $\\ell_2$-Boosting Overfit Benignly? High-Dimensional Risk Asymptotics and the $\\ell_1$ Implicit Bias","display_name":"When Does $\\ell_2$-Boosting Overfit Benignly? High-Dimensional Risk Asymptotics and the $\\ell_1$ Implicit Bias","publication_year":2026,"publication_date":"2026-05-07","ids":{"openalex":"https://openalex.org/W7160620685","doi":"https://doi.org/10.48550/arxiv.2605.06314"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2605.06314","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.06314","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":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.2605.06314","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5135681957","display_name":"Ye Su","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Su, Ye","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135694365","display_name":"Jian Li","orcid":"https://orcid.org/0009-0003-3390-6335"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Jian","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5135659090","display_name":"Yong Liu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Yong","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/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.2029999941587448,"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/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.2029999941587448,"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/T11830","display_name":"Point processes and geometric inequalities","score":0.1509999930858612,"subfield":{"id":"https://openalex.org/subfields/2604","display_name":"Applied Mathematics"},"field":{"id":"https://openalex.org/fields/26","display_name":"Mathematics"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10136","display_name":"Statistical Methods and Inference","score":0.12520000338554382,"subfield":{"id":"https://openalex.org/subfields/2613","display_name":"Statistics and Probability"},"field":{"id":"https://openalex.org/fields/26","display_name":"Mathematics"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/overfitting","display_name":"Overfitting","score":0.7035999894142151},{"id":"https://openalex.org/keywords/logarithm","display_name":"Logarithm","score":0.4812000095844269},{"id":"https://openalex.org/keywords/gaussian","display_name":"Gaussian","score":0.3986000120639801},{"id":"https://openalex.org/keywords/bounded-function","display_name":"Bounded function","score":0.38940000534057617},{"id":"https://openalex.org/keywords/minimax","display_name":"Minimax","score":0.36980000138282776},{"id":"https://openalex.org/keywords/eigenvalues-and-eigenvectors","display_name":"Eigenvalues and eigenvectors","score":0.36059999465942383},{"id":"https://openalex.org/keywords/model-selection","display_name":"Model selection","score":0.34869998693466187},{"id":"https://openalex.org/keywords/gaussian-noise","display_name":"Gaussian noise","score":0.3424000144004822}],"concepts":[{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.7555999755859375},{"id":"https://openalex.org/C22019652","wikidata":"https://www.wikidata.org/wiki/Q331309","display_name":"Overfitting","level":3,"score":0.7035999894142151},{"id":"https://openalex.org/C28826006","wikidata":"https://www.wikidata.org/wiki/Q33521","display_name":"Applied mathematics","level":1,"score":0.5134999752044678},{"id":"https://openalex.org/C39927690","wikidata":"https://www.wikidata.org/wiki/Q11197","display_name":"Logarithm","level":2,"score":0.4812000095844269},{"id":"https://openalex.org/C163716315","wikidata":"https://www.wikidata.org/wiki/Q901177","display_name":"Gaussian","level":2,"score":0.3986000120639801},{"id":"https://openalex.org/C34388435","wikidata":"https://www.wikidata.org/wiki/Q2267362","display_name":"Bounded function","level":2,"score":0.38940000534057617},{"id":"https://openalex.org/C149728462","wikidata":"https://www.wikidata.org/wiki/Q751319","display_name":"Minimax","level":2,"score":0.36980000138282776},{"id":"https://openalex.org/C158693339","wikidata":"https://www.wikidata.org/wiki/Q190524","display_name":"Eigenvalues and eigenvectors","level":2,"score":0.36059999465942383},{"id":"https://openalex.org/C93959086","wikidata":"https://www.wikidata.org/wiki/Q6888345","display_name":"Model selection","level":2,"score":0.34869998693466187},{"id":"https://openalex.org/C4199805","wikidata":"https://www.wikidata.org/wiki/Q2725903","display_name":"Gaussian noise","level":2,"score":0.3424000144004822},{"id":"https://openalex.org/C37616216","wikidata":"https://www.wikidata.org/wiki/Q3218363","display_name":"Lasso (programming language)","level":2,"score":0.3122999966144562},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.29089999198913574},{"id":"https://openalex.org/C196083921","wikidata":"https://www.wikidata.org/wiki/Q7915758","display_name":"Variance (accounting)","level":2,"score":0.2856000065803528},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.28540000319480896},{"id":"https://openalex.org/C112680207","wikidata":"https://www.wikidata.org/wiki/Q714886","display_name":"Regular polygon","level":2,"score":0.2768999934196472},{"id":"https://openalex.org/C81917197","wikidata":"https://www.wikidata.org/wiki/Q628760","display_name":"Selection (genetic algorithm)","level":2,"score":0.27489998936653137},{"id":"https://openalex.org/C62799726","wikidata":"https://www.wikidata.org/wiki/Q190056","display_name":"Hilbert space","level":2,"score":0.274399995803833},{"id":"https://openalex.org/C61326573","wikidata":"https://www.wikidata.org/wiki/Q1496376","display_name":"Gaussian process","level":3,"score":0.26969999074935913},{"id":"https://openalex.org/C77553402","wikidata":"https://www.wikidata.org/wiki/Q13222579","display_name":"Upper and lower bounds","level":2,"score":0.2676999866962433},{"id":"https://openalex.org/C139912040","wikidata":"https://www.wikidata.org/wiki/Q937378","display_name":"Logarithmic scale","level":2,"score":0.26759999990463257},{"id":"https://openalex.org/C2776029896","wikidata":"https://www.wikidata.org/wiki/Q3935810","display_name":"Relaxation (psychology)","level":2,"score":0.25200000405311584}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2605.06314","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.06314","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":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2605.06314","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.06314","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":false,"raw_source_name":null,"raw_type":"Preprint"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Benign":[0],"overfitting":[1,89],"is":[2,171],"well-characterized":[3],"in":[4,126,178],"$\\ell_2$":[5,179],"geometries,":[6],"but":[7,163],"its":[8],"behavior":[9],"under":[10,204],"the":[11,25,45,59,76,92,104,127,131,154,174,188,193,210,216],"$\\ell_1$":[12,78],"implicit":[13],"bias":[14],"of":[15,28,48,68,129,192],"greedy":[16,95],"ensembles":[17],"remains":[18,135],"challenging.":[19],"The":[20],"analytical":[21],"barrier":[22],"stems":[23],"from":[24],"non-linear":[26],"coupling":[27,58],"coordinate":[29],"selection":[30,96],"thresholds,":[31],"which":[32,170],"invalidates":[33],"standard":[34],"spectral":[35],"resolvent":[36],"tools.":[37],"To":[38,181],"isolate":[39],"this":[40,121,183],"algorithmic":[41],"bias,":[42],"we":[43,73,85,186],"characterize":[44],"high-dimensional":[46],"risk":[47,155],"continuous-time":[49],"$\\ell_2$-Boosting":[50],"over":[51],"$p$":[52],"features":[53],"and":[54,103,146,214],"$n$":[55],"samples.":[56],"By":[57],"Convex":[60],"Gaussian":[61,71],"Minimax":[62],"Theorem":[63],"with":[64,142],"delicate":[65],"asymptotic":[66],"expansions":[67],"double-sided":[69],"truncated":[70],"moments,":[72],"analytically":[74],"resolve":[75],"non-smooth":[77,189],"interpolant.":[79],"Under":[80],"an":[81,136],"isotropic":[82],"pure-noise":[83],"model,":[84],"prove":[86],"that":[87,119],"benign":[88],"fails":[90],"at":[91,108,165],"linear":[93,175],"rate:":[94],"localizes":[97],"noise":[98,114],"into":[99],"sparse":[100],"active":[101],"sets,":[102],"excess":[105],"variance":[106,115],"decays":[107],"a":[109,166,198,205],"logarithmic":[110,167],"rate":[111,168,220],"$\u0398(\u03c3^2/\\log(p/n))$":[112],"for":[113,221],"$\u03c3^2$.":[116],"We":[117],"remark":[118],"while":[120],"localization":[122],"mechanism":[123],"should":[124],"persist":[125],"presence":[128],"signals,":[130],"exact":[132],"signal-noise":[133],"decomposition":[134],"open":[137],"problem.":[138],"For":[139],"spiked-isotropic":[140],"designs":[141],"$k^*$":[143],"head":[144],"eigenvalues":[145],"$r_2":[147],"=":[148],"p":[149],"-":[150],"k^*$":[151],"tail":[152],"dimensions,":[153],"converges":[156],"to":[157],"zero":[158],"when":[159],"$r_{2}":[160],"\\gg":[161],"n$,":[162],"only":[164],"$\u0398(\u03c3^2/\\log(r_2/n))$,":[169],"slower":[172],"than":[173],"decay":[176],"observed":[177],"geometries.":[180],"avoid":[182],"slow":[184],"convergence,":[185],"analyze":[187],"subdifferential":[190],"dynamics":[191],"boosting":[194],"flow.":[195],"This":[196],"yields":[197],"tuning-free":[199],"early":[200],"stopping":[201],"rule":[202],"that,":[203],"bounded":[206],"$\\ell_1$-path":[207],"condition,":[208],"recovers":[209],"Lasso":[211],"basic":[212],"inequality":[213],"attains":[215],"minimax-optimal":[217],"empirical":[218],"prediction":[219],"$\\ell_1$-bounded":[222],"signals.":[223]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-05-09T00:00:00"}