{"id":"https://openalex.org/W7115603752","doi":"https://doi.org/10.48550/arxiv.2512.11089","title":"TPV: Parameter Perturbations Through the Lens of Test Prediction Variance","display_name":"TPV: Parameter Perturbations Through the Lens of Test Prediction Variance","publication_year":2025,"publication_date":"2025-12-11","ids":{"openalex":"https://openalex.org/W7115603752","doi":"https://doi.org/10.48550/arxiv.2512.11089"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2512.11089","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2512.11089","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":"article"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2512.11089","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":null,"display_name":"Arpit, Devansh","orcid":null},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Arpit, Devansh","raw_affiliation_strings":[],"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":1,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.2799000144004822,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.2799000144004822,"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/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.16840000450611115,"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":0.1467999964952469,"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/generalization","display_name":"Generalization","score":0.7576000094413757},{"id":"https://openalex.org/keywords/perturbation","display_name":"Perturbation (astronomy)","score":0.6974999904632568},{"id":"https://openalex.org/keywords/variance","display_name":"Variance (accounting)","score":0.6134999990463257},{"id":"https://openalex.org/keywords/stability","display_name":"Stability (learning theory)","score":0.5835999846458435},{"id":"https://openalex.org/keywords/metric","display_name":"Metric (unit)","score":0.5807999968528748},{"id":"https://openalex.org/keywords/set","display_name":"Set (abstract data type)","score":0.46779999136924744},{"id":"https://openalex.org/keywords/variance-components","display_name":"Variance components","score":0.36739999055862427},{"id":"https://openalex.org/keywords/training-set","display_name":"Training set","score":0.36149999499320984},{"id":"https://openalex.org/keywords/sensitivity","display_name":"Sensitivity (control systems)","score":0.3578999936580658}],"concepts":[{"id":"https://openalex.org/C177148314","wikidata":"https://www.wikidata.org/wiki/Q170084","display_name":"Generalization","level":2,"score":0.7576000094413757},{"id":"https://openalex.org/C177918212","wikidata":"https://www.wikidata.org/wiki/Q803623","display_name":"Perturbation (astronomy)","level":2,"score":0.6974999904632568},{"id":"https://openalex.org/C196083921","wikidata":"https://www.wikidata.org/wiki/Q7915758","display_name":"Variance (accounting)","level":2,"score":0.6134999990463257},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.5870000123977661},{"id":"https://openalex.org/C112972136","wikidata":"https://www.wikidata.org/wiki/Q7595718","display_name":"Stability (learning theory)","level":2,"score":0.5835999846458435},{"id":"https://openalex.org/C176217482","wikidata":"https://www.wikidata.org/wiki/Q860554","display_name":"Metric (unit)","level":2,"score":0.5807999968528748},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.46779999136924744},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.44339999556541443},{"id":"https://openalex.org/C28826006","wikidata":"https://www.wikidata.org/wiki/Q33521","display_name":"Applied mathematics","level":1,"score":0.3749000132083893},{"id":"https://openalex.org/C3018076075","wikidata":"https://www.wikidata.org/wiki/Q1826427","display_name":"Variance components","level":2,"score":0.36739999055862427},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.36149999499320984},{"id":"https://openalex.org/C21200559","wikidata":"https://www.wikidata.org/wiki/Q7451068","display_name":"Sensitivity (control systems)","level":2,"score":0.3578999936580658},{"id":"https://openalex.org/C169903167","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Test set","level":2,"score":0.3292999863624573},{"id":"https://openalex.org/C2780898871","wikidata":"https://www.wikidata.org/wiki/Q860554","display_name":"Performance metric","level":2,"score":0.3264000117778778},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.31779998540878296},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.314300000667572},{"id":"https://openalex.org/C204323151","wikidata":"https://www.wikidata.org/wiki/Q905424","display_name":"Range (aeronautics)","level":2,"score":0.30799999833106995},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.29750001430511475},{"id":"https://openalex.org/C75291252","wikidata":"https://www.wikidata.org/wiki/Q1315756","display_name":"TRACE (psycholinguistics)","level":2,"score":0.2948000133037567},{"id":"https://openalex.org/C167085575","wikidata":"https://www.wikidata.org/wiki/Q6803654","display_name":"Mean squared prediction error","level":2,"score":0.2840000092983246},{"id":"https://openalex.org/C2781238097","wikidata":"https://www.wikidata.org/wiki/Q175026","display_name":"Object (grammar)","level":2,"score":0.27300000190734863},{"id":"https://openalex.org/C167928553","wikidata":"https://www.wikidata.org/wiki/Q1376021","display_name":"Estimation theory","level":2,"score":0.26980000734329224},{"id":"https://openalex.org/C134306372","wikidata":"https://www.wikidata.org/wiki/Q7754","display_name":"Mathematical analysis","level":1,"score":0.2596000134944916},{"id":"https://openalex.org/C2777303404","wikidata":"https://www.wikidata.org/wiki/Q759757","display_name":"Convergence (economics)","level":2,"score":0.25850000977516174},{"id":"https://openalex.org/C3019719930","wikidata":"https://www.wikidata.org/wiki/Q3910099","display_name":"Predictive value","level":2,"score":0.25839999318122864},{"id":"https://openalex.org/C15336307","wikidata":"https://www.wikidata.org/wiki/Q1766051","display_name":"Lens (geology)","level":2,"score":0.25450000166893005},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.2515000104904175}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2512.11089","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2512.11089","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":"article"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2512.11089","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2512.11089","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":"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":{"We":[0],"identify":[1],"test":[2,144],"prediction":[3,103],"variance":[4,104],"(TPV)--":[5],"the":[6,43,46,50,86,95,99],"first-order":[7],"sensitivity":[8],"of":[9,45,58],"model":[10,48],"outputs":[11],"to":[12,72,90],"parameter":[13,59,107],"perturbations":[14,60,71,108],"around":[15],"a":[16,20,35,55,76,127,148],"trained":[17,47],"solution--":[18],"as":[19,147],"unifying":[21],"quantity":[22],"that":[23,82,102],"links":[24],"several":[25],"classical":[26],"observations":[27],"about":[28],"generalization":[29,122],"in":[30,94],"deep":[31],"networks.":[32,138],"TPV":[33,83,125,140],"is":[34,119],"fully":[36],"label-free":[37],"object":[38],"whose":[39],"trace":[40],"form":[41],"separates":[42],"geometry":[44],"from":[49,112,121],"specific":[51],"perturbation":[52],"mechanism,":[53],"allowing":[54],"broad":[56],"family":[57],"like":[61],"SGD":[62],"noise,":[63,65,67],"label":[64],"finite-precision":[66],"and":[68,116,132],"other":[69],"post-training":[70],"be":[73,110],"analyzed":[74],"under":[75,105],"single":[77],"framework.":[78],"Theoretically,":[79],"we":[80],"show":[81],"estimated":[84],"on":[85],"training":[87,113],"set":[88],"converges":[89],"its":[91],"test-set":[92],"value":[93],"overparameterized":[96],"limit,":[97],"providing":[98],"first":[100],"result":[101],"local":[106],"can":[109],"inferred":[111],"inputs":[114],"alone,":[115],"this":[117],"stability":[118,129],"decoupled":[120],"performance.":[123],"Empirically,":[124],"exhibits":[126],"striking":[128],"across":[130],"datasets":[131],"architectures":[133],"even":[134],"for":[135,153],"extremely":[136],"narrow":[137],"Further,":[139],"correlates":[141],"well":[142],"with":[143],"loss,":[145],"serving":[146],"training-set":[149],"based":[150],"predictive":[151],"metric":[152],"generalization.":[154],"Code":[155],"available":[156],"at":[157],"github.com/devansharpit/TPV.":[158]},"counts_by_year":[],"updated_date":"2026-01-24T23:23:39.755997","created_date":"2025-12-16T00:00:00"}