{"id":"https://openalex.org/W7124144853","doi":"https://doi.org/10.48550/arxiv.2601.08316","title":"Deep Exploration of Epoch-wise Double Descent in Noisy Data: Signal Separation, Large Activation, and Benign Overfitting","display_name":"Deep Exploration of Epoch-wise Double Descent in Noisy Data: Signal Separation, Large Activation, and Benign Overfitting","publication_year":2026,"publication_date":"2026-01-13","ids":{"openalex":"https://openalex.org/W7124144853","doi":"https://doi.org/10.48550/arxiv.2601.08316"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2601.08316","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2601.08316","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"article"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2601.08316","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5123010627","display_name":"Tomoki Kubo","orcid":null},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Kubo, Tomoki","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5123012346","display_name":"Ryuken Uda","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Uda, Ryuken","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5122996651","display_name":"Yusuke Iida","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Iida, Yusuke","raw_affiliation_strings":[],"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":3,"corresponding_author_ids":["https://openalex.org/A5123010627"],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":1,"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/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.1973000019788742,"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"}},"topics":[{"id":"https://openalex.org/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.1973000019788742,"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/T12535","display_name":"Machine Learning and Data Classification","score":0.08479999750852585,"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/T10320","display_name":"Neural Networks and Applications","score":0.07559999823570251,"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/overfitting","display_name":"Overfitting","score":0.9487000107765198},{"id":"https://openalex.org/keywords/generalization","display_name":"Generalization","score":0.6710000038146973},{"id":"https://openalex.org/keywords/deep-learning","display_name":"Deep learning","score":0.6172999739646912},{"id":"https://openalex.org/keywords/noise","display_name":"Noise (video)","score":0.566100001335144},{"id":"https://openalex.org/keywords/signal","display_name":"SIGNAL (programming language)","score":0.5655999779701233},{"id":"https://openalex.org/keywords/gradient-descent","display_name":"Gradient descent","score":0.5565999746322632},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.5522000193595886},{"id":"https://openalex.org/keywords/stochastic-gradient-descent","display_name":"Stochastic gradient descent","score":0.5367000102996826}],"concepts":[{"id":"https://openalex.org/C22019652","wikidata":"https://www.wikidata.org/wiki/Q331309","display_name":"Overfitting","level":3,"score":0.9487000107765198},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.6890000104904175},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6816999912261963},{"id":"https://openalex.org/C177148314","wikidata":"https://www.wikidata.org/wiki/Q170084","display_name":"Generalization","level":2,"score":0.6710000038146973},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.6172999739646912},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.566100001335144},{"id":"https://openalex.org/C2779843651","wikidata":"https://www.wikidata.org/wiki/Q7390335","display_name":"SIGNAL (programming language)","level":2,"score":0.5655999779701233},{"id":"https://openalex.org/C153258448","wikidata":"https://www.wikidata.org/wiki/Q1199743","display_name":"Gradient descent","level":3,"score":0.5565999746322632},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.5522000193595886},{"id":"https://openalex.org/C206688291","wikidata":"https://www.wikidata.org/wiki/Q7617819","display_name":"Stochastic gradient descent","level":3,"score":0.5367000102996826},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.5123000144958496},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.5058000087738037},{"id":"https://openalex.org/C2984842247","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep neural networks","level":3,"score":0.4120999872684479},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.37229999899864197},{"id":"https://openalex.org/C2781170535","wikidata":"https://www.wikidata.org/wiki/Q30587856","display_name":"Noisy data","level":2,"score":0.3666999936103821},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3416000008583069},{"id":"https://openalex.org/C112633086","wikidata":"https://www.wikidata.org/wiki/Q381287","display_name":"White noise","level":2,"score":0.3237000107765198},{"id":"https://openalex.org/C2777303404","wikidata":"https://www.wikidata.org/wiki/Q759757","display_name":"Convergence (economics)","level":2,"score":0.3208000063896179},{"id":"https://openalex.org/C16910744","wikidata":"https://www.wikidata.org/wiki/Q7705759","display_name":"Test data","level":2,"score":0.2996000051498413},{"id":"https://openalex.org/C5465570","wikidata":"https://www.wikidata.org/wiki/Q5326898","display_name":"Early stopping","level":3,"score":0.2793999910354614},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.2687999904155731},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.25060001015663147}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2601.08316","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2601.08316","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"article"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2601.08316","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2601.08316","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"article"},"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":{"Deep":[0],"double":[1,21,107,206,224],"descent":[2,108],"is":[3,24,163],"one":[4],"of":[5,13,37,44,75,183,204,217],"the":[6,10,35,51,60,72,90,106,140,155,200,215],"key":[7,202],"phenomena":[8,203],"underlying":[9],"generalization":[11,26],"capability":[12],"deep":[14,223],"learning":[15,126],"models.":[16],"In":[17],"this":[18,87,138,161],"study,":[19],"epoch-wise":[20,73],"descent,":[22],"which":[23],"delayed":[25],"following":[27],"overfitting,":[28],"was":[29],"empirically":[30],"investigated":[31],"by":[32],"focusing":[33],"on":[34,50,95],"evolution":[36],"internal":[38,76,130],"structures.":[39],"Fully":[40],"connected":[41],"neural":[42],"networks":[43],"three":[45],"different":[46],"sizes":[47],"were":[48,78,84,119],"trained":[49],"CIFAR-10":[52],"dataset":[53],"with":[54,179,187,192],"30%":[55],"label":[56],"noise.":[57],"By":[58],"decomposing":[59],"loss":[61],"curves":[62],"into":[63],"signal":[64],"contributions":[65],"from":[66,86],"clean":[67,122],"and":[68,124,169,177,210,213],"noisy":[69,102,117,145],"training":[70,103],"data,":[71,123],"evolutions":[74],"signals":[77],"analyzed":[79],"separately.":[80],"Three":[81],"main":[82],"findings":[83,197],"obtained":[85],"analysis.":[88],"First,":[89],"model":[91,141],"achieved":[92],"strong":[93],"re-generalization":[94],"test":[96],"data":[97,104,118],"even":[98],"after":[99,121],"perfectly":[100],"fitting":[101],"during":[105],"phase,":[109],"corresponding":[110,129],"to":[111,142],"a":[112,148,218],"\"benign":[113,208],"overfitting\"":[114],"state.":[115],"Second,":[116],"learned":[120],"as":[125,165],"progressed,":[127],"their":[128],"activations":[131],"became":[132],"increasingly":[133],"separated":[134],"in":[135,154,172],"outer":[136],"layers;":[137],"enabled":[139],"overfit":[143],"only":[144],"data.":[146],"Third,":[147],"single,":[149],"very":[150],"large":[151,174,184],"activation":[152,185],"emerged":[153],"shallow":[156],"layer":[157],"across":[158],"all":[159],"models;":[160],"phenomenon":[162],"referred":[164],"\"outliers,\"":[166],"\"massive":[167],"activa-tions,\"":[168],"\"super":[170],"activations\"":[171],"recent":[173,201],"language":[175],"models":[176],"evolves":[178],"re-generalization.":[180],"The":[181],"magnitude":[182],"correlated":[186],"input":[188],"patterns":[189],"but":[190],"not":[191],"output":[193],"patterns.":[194],"These":[195],"empirical":[196],"directly":[198],"link":[199],"\"deep":[205],"descent,\"":[207],"overfitting,\"":[209],"\"large":[211],"activation\",":[212],"support":[214],"proposal":[216],"novel":[219],"scenario":[220],"for":[221],"understanding":[222],"descent.":[225]},"counts_by_year":[{"year":2026,"cited_by_count":1}],"updated_date":"2026-03-25T14:56:36.534964","created_date":"2026-01-15T00:00:00"}