{"id":"https://openalex.org/W7161681662","doi":"https://doi.org/10.48550/arxiv.2605.18180","title":"Canonical Regularisation of Wide Feature-Learning Neural Networks","display_name":"Canonical Regularisation of Wide Feature-Learning Neural Networks","publication_year":2026,"publication_date":"2026-05-18","ids":{"openalex":"https://openalex.org/W7161681662","doi":"https://doi.org/10.48550/arxiv.2605.18180"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2605.18180","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.18180","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.2605.18180","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5116167651","display_name":"George Whittle","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Whittle, George","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136499836","display_name":"Pranav Vaidhyanathan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Vaidhyanathan, Pranav","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5001672680","display_name":"Juliusz Ziomek","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ziomek, Juliusz","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5020583292","display_name":"Natalia Ares","orcid":"https://orcid.org/0000-0003-2588-6322"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ares, Natalia","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5136479340","display_name":"Maike A. Osborne","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Osborne, Maike A.","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":0.6647999882698059,"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":0.6647999882698059,"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/T12814","display_name":"Gaussian Processes and Bayesian Inference","score":0.131400004029274,"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/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.019700000062584877,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/geodesic","display_name":"Geodesic","score":0.7092000246047974},{"id":"https://openalex.org/keywords/ridge","display_name":"Ridge","score":0.6037999987602234},{"id":"https://openalex.org/keywords/balanced-flow","display_name":"Balanced flow","score":0.5263000130653381},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.47360000014305115},{"id":"https://openalex.org/keywords/kernel","display_name":"Kernel (algebra)","score":0.4131999909877777},{"id":"https://openalex.org/keywords/monotonic-function","display_name":"Monotonic function","score":0.3921000063419342},{"id":"https://openalex.org/keywords/flow","display_name":"Flow (mathematics)","score":0.37770000100135803},{"id":"https://openalex.org/keywords/gaussian","display_name":"Gaussian","score":0.376800000667572},{"id":"https://openalex.org/keywords/deep-neural-networks","display_name":"Deep neural networks","score":0.3296000063419342}],"concepts":[{"id":"https://openalex.org/C165818556","wikidata":"https://www.wikidata.org/wiki/Q213488","display_name":"Geodesic","level":2,"score":0.7092000246047974},{"id":"https://openalex.org/C32277403","wikidata":"https://www.wikidata.org/wiki/Q740445","display_name":"Ridge","level":2,"score":0.6037999987602234},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.5565999746322632},{"id":"https://openalex.org/C167879884","wikidata":"https://www.wikidata.org/wiki/Q727568","display_name":"Balanced flow","level":2,"score":0.5263000130653381},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.47360000014305115},{"id":"https://openalex.org/C74193536","wikidata":"https://www.wikidata.org/wiki/Q574844","display_name":"Kernel (algebra)","level":2,"score":0.4131999909877777},{"id":"https://openalex.org/C72169020","wikidata":"https://www.wikidata.org/wiki/Q194404","display_name":"Monotonic function","level":2,"score":0.3921000063419342},{"id":"https://openalex.org/C38349280","wikidata":"https://www.wikidata.org/wiki/Q1434290","display_name":"Flow (mathematics)","level":2,"score":0.37770000100135803},{"id":"https://openalex.org/C163716315","wikidata":"https://www.wikidata.org/wiki/Q901177","display_name":"Gaussian","level":2,"score":0.376800000667572},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3504999876022339},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3368000090122223},{"id":"https://openalex.org/C2984842247","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep neural networks","level":3,"score":0.3296000063419342},{"id":"https://openalex.org/C151201525","wikidata":"https://www.wikidata.org/wiki/Q177239","display_name":"Limit (mathematics)","level":2,"score":0.3240000009536743},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.3147999942302704},{"id":"https://openalex.org/C91229774","wikidata":"https://www.wikidata.org/wiki/Q193885","display_name":"Infinitesimal","level":2,"score":0.31360000371932983},{"id":"https://openalex.org/C28826006","wikidata":"https://www.wikidata.org/wiki/Q33521","display_name":"Applied mathematics","level":1,"score":0.3100999891757965},{"id":"https://openalex.org/C121864883","wikidata":"https://www.wikidata.org/wiki/Q677916","display_name":"Statistical physics","level":1,"score":0.3057999908924103},{"id":"https://openalex.org/C192939610","wikidata":"https://www.wikidata.org/wiki/Q188444","display_name":"Differential geometry","level":2,"score":0.2849999964237213},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.2838999927043915},{"id":"https://openalex.org/C80899671","wikidata":"https://www.wikidata.org/wiki/Q1304193","display_name":"Vertex (graph theory)","level":3,"score":0.27639999985694885},{"id":"https://openalex.org/C189950617","wikidata":"https://www.wikidata.org/wiki/Q937228","display_name":"Property (philosophy)","level":2,"score":0.271699994802475},{"id":"https://openalex.org/C73313986","wikidata":"https://www.wikidata.org/wiki/Q355386","display_name":"Luminance","level":2,"score":0.26930001378059387},{"id":"https://openalex.org/C204707403","wikidata":"https://www.wikidata.org/wiki/Q1152398","display_name":"Canonical form","level":2,"score":0.26179999113082886},{"id":"https://openalex.org/C134306372","wikidata":"https://www.wikidata.org/wiki/Q7754","display_name":"Mathematical analysis","level":1,"score":0.26159998774528503},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.2587999999523163},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.25619998574256897},{"id":"https://openalex.org/C2524010","wikidata":"https://www.wikidata.org/wiki/Q8087","display_name":"Geometry","level":1,"score":0.2531000077724457},{"id":"https://openalex.org/C204241405","wikidata":"https://www.wikidata.org/wiki/Q461499","display_name":"Transformation (genetics)","level":3,"score":0.2529999911785126},{"id":"https://openalex.org/C109546454","wikidata":"https://www.wikidata.org/wiki/Q3798604","display_name":"Information geometry","level":4,"score":0.25209999084472656}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2605.18180","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.18180","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.2605.18180","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.18180","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":{"Wide":[0],"neural":[1],"networks":[2,51,119],"in":[3,48,90,95,145],"the":[4,33,55,63,70,76,96,106,110,121,131,146,153,158,174,180,190,227],"feature-learning":[5,91,154,162,175],"regime":[6,50,92],"drive":[7],"modern":[8],"deep":[9,213],"learning,":[10],"and":[11,35,68,139,149,218,237],"yet":[12,26],"they":[13],"remain":[14],"far":[15],"less":[16],"studied":[17],"than":[18],"their":[19],"kernel-regime":[20],"counterparts.":[21],"We":[22,126],"consider":[23],"a":[24,113,135,185,196,204,212],"critical":[25],"under-explored":[27],"difference":[28],"between":[29,215],"these":[30],"two":[31],"regimes:":[32],"regulariser":[34,133],"prior":[36,123,183],"implied":[37],"by":[38,129],"gradient":[39,59,88],"flow":[40,60,89],"training.":[41,81],"This":[42],"canonical":[43,132,181,219],"regularisation":[44,86,220],"property":[45],"is":[46,124,184],"well-studied":[47],"kernel":[49,147],"--":[52,67],"of":[53,78,99,109,161,229],"all":[54],"infinite":[56],"global":[57],"minima,":[58],"selects":[61],"exactly":[62],"vanishing":[64,100],"ridge":[65,85,104,144,167,172,202],"solution":[66],"underpins":[69],"celebrated":[71],"NN-GP":[72],"correspondence,":[73],"precisely":[74],"allowing":[75],"modelling":[77],"noise":[79],"during":[80],"However,":[82],"we":[83,164,178,199,225],"prove":[84,179],"biases":[87],"networks,":[93,163],"even":[94],"infinitesimal":[97],"limit":[98],"regularisation.":[101],"Over":[102],"training,":[103],"distorts":[105],"inductive":[107],"bias":[108],"network,":[111],"with":[112],"particular":[114],"damage":[115],"done":[116],"to":[117,152,173,208],"pretrained":[118],"where":[120],"implicit":[122],"informative.":[125],"resolve":[127],"this":[128],"axiomatising":[130],"as":[134,203],"regime-agnostic":[136],"function-space":[137,182],"energy":[138],"lift,":[140],"which":[141],"uniquely":[142],"identifies":[143],"regime,":[148],"crucially":[150],"generalises":[151],"regime.":[155,176],"By":[156],"studying":[157],"Riemannian":[159,186],"geometry":[160],"derive":[165],"geodesic":[166,209],"from":[168],"our":[169,230],"framework,":[170],"generalising":[171,189],"Correspondingly,":[177],"Gibbs":[187],"Process,":[188],"more":[191],"familiar":[192],"Gaussian":[193],"Process.":[194],"As":[195],"practical":[197],"contribution,":[198],"propose":[200],"arc":[201],"minimax-robust,":[205],"scalable":[206],"surrogate":[207],"ridge,":[210],"revealing":[211],"relationship":[214],"early":[216],"stopping":[217],"across":[221],"learning":[222],"regimes.":[223],"Finally,":[224],"demonstrate":[226],"consequences":[228],"theory":[231],"empirically":[232],"on":[233],"both":[234],"image":[235],"processing":[236],"NLP":[238],"transfer-learning":[239],"problems.":[240]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-05-20T00:00:00"}