FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3117245434> ?p ?o ?g. }

• W3117245434 abstract "The presence of label noise often misleads the training of deep neural networks. Departing from the recent literature which largely assumes the label noise rate is only determined by the true label class, the errors in human-annotated labels are more likely to be dependent on the difficulty levels of tasks, resulting in settings with instance-dependent label noise. We first provide evidences that the heterogeneous instance-dependent label noise is effectively down-weighting the examples with higher noise rates in a non-uniform way and thus causes imbalances, rendering the strategy of directly applying methods for class-dependent label noise questionable. Built on a recent work peer loss [24], we then propose and study the potentials of a second-order approach that leverages the estimation of several covariance terms defined between the instance-dependent noise rates and the Bayes optimal label. We show that this set of second-order statistics successfully captures the induced imbalances. We further proceed to show that with the help of the estimated second-order statistics, we identify a new loss function whose expected risk of a classifier under instance-dependent label noise is equivalent to a new problem with only class-dependent label noise. This fact allows us to apply existing solutions to handle this better-studied setting. We provide an efficient procedure to estimate these second-order statistics without accessing either ground truth labels or prior knowledge of the noise rates. Experiments on CIFAR10 and CIFAR100 with synthetic instance-dependent label noise and Clothing1M with real-world human label noise verify our approach. Our implementation is available at https://github.com/UCSC-REAL/CAL." @default.
• W3117245434 created "2021-01-05" @default.
• W3117245434 creator A5023363049 @default.
• W3117245434 creator A5025441627 @default.
• W3117245434 creator A5065250332 @default.
• W3117245434 date "2020-12-22" @default.
• W3117245434 modified "2023-09-25" @default.
• W3117245434 title "A Second-Order Approach to Learning with Instance-Dependent Label Noise" @default.
• W3117245434 cites W1514928307 @default.
• W3117245434 cites W1921293667 @default.
• W3117245434 cites W1982032418 @default.
• W3117245434 cites W2102348129 @default.
• W3117245434 cites W2112483442 @default.
• W3117245434 cites W2113290770 @default.
• W3117245434 cites W2490901831 @default.
• W3117245434 cites W2577784528 @default.
• W3117245434 cites W2592335154 @default.
• W3117245434 cites W2810089099 @default.
• W3117245434 cites W2811380766 @default.
• W3117245434 cites W2909793444 @default.
• W3117245434 cites W2950220847 @default.
• W3117245434 cites W2962762068 @default.
• W3117245434 cites W2963031075 @default.
• W3117245434 cites W2963081269 @default.
• W3117245434 cites W2963399829 @default.
• W3117245434 cites W2963697299 @default.
• W3117245434 cites W2963735582 @default.
• W3117245434 cites W2963759070 @default.
• W3117245434 cites W2964096266 @default.
• W3117245434 cites W2964292098 @default.
• W3117245434 cites W2970038028 @default.
• W3117245434 cites W2970308742 @default.
• W3117245434 cites W2971193222 @default.
• W3117245434 cites W2981873476 @default.
• W3117245434 cites W2990019157 @default.
• W3117245434 cites W2995624272 @default.
• W3117245434 cites W2998375045 @default.
• W3117245434 cites W3005871659 @default.
• W3117245434 cites W3034185248 @default.
• W3117245434 cites W3034711955 @default.
• W3117245434 cites W3035314656 @default.
• W3117245434 cites W3035702670 @default.
• W3117245434 cites W3035753488 @default.
• W3117245434 cites W3092841986 @default.
• W3117245434 cites W3102433174 @default.
• W3117245434 cites W3102643535 @default.
• W3117245434 cites W3103846556 @default.
• W3117245434 cites W3104255659 @default.
• W3117245434 cites W3110483780 @default.
• W3117245434 cites W3118608800 @default.
• W3117245434 cites W3118807731 @default.
• W3117245434 cites W3126748415 @default.
• W3117245434 cites W3127098609 @default.
• W3117245434 cites W3129100675 @default.
• W3117245434 cites W3136901775 @default.
• W3117245434 doi "https://doi.org/10.48550/arxiv.2012.11854" @default.
• W3117245434 hasPublicationYear "2020" @default.
• W3117245434 type Work @default.
• W3117245434 sameAs 3117245434 @default.
• W3117245434 citedByCount "3" @default.
• W3117245434 countsByYear W31172454342020 @default.
• W3117245434 countsByYear W31172454342021 @default.
• W3117245434 crossrefType "posted-content" @default.
• W3117245434 hasAuthorship W3117245434A5023363049 @default.
• W3117245434 hasAuthorship W3117245434A5025441627 @default.
• W3117245434 hasAuthorship W3117245434A5065250332 @default.
• W3117245434 hasBestOaLocation W31172454341 @default.
• W3117245434 hasConcept C107673813 @default.
• W3117245434 hasConcept C115961682 @default.
• W3117245434 hasConcept C119857082 @default.
• W3117245434 hasConcept C126838900 @default.
• W3117245434 hasConcept C153180895 @default.
• W3117245434 hasConcept C154945302 @default.
• W3117245434 hasConcept C183115368 @default.
• W3117245434 hasConcept C207201462 @default.
• W3117245434 hasConcept C41008148 @default.
• W3117245434 hasConcept C71924100 @default.
• W3117245434 hasConcept C95623464 @default.
• W3117245434 hasConcept C99498987 @default.
• W3117245434 hasConceptScore W3117245434C107673813 @default.
• W3117245434 hasConceptScore W3117245434C115961682 @default.
• W3117245434 hasConceptScore W3117245434C119857082 @default.
• W3117245434 hasConceptScore W3117245434C126838900 @default.
• W3117245434 hasConceptScore W3117245434C153180895 @default.
• W3117245434 hasConceptScore W3117245434C154945302 @default.
• W3117245434 hasConceptScore W3117245434C183115368 @default.
• W3117245434 hasConceptScore W3117245434C207201462 @default.
• W3117245434 hasConceptScore W3117245434C41008148 @default.
• W3117245434 hasConceptScore W3117245434C71924100 @default.
• W3117245434 hasConceptScore W3117245434C95623464 @default.
• W3117245434 hasConceptScore W3117245434C99498987 @default.
• W3117245434 hasLocation W31172454341 @default.
• W3117245434 hasOpenAccess W3117245434 @default.
• W3117245434 hasPrimaryLocation W31172454341 @default.
• W3117245434 hasRelatedWork W2001652754 @default.
• W3117245434 hasRelatedWork W2069189854 @default.
• W3117245434 hasRelatedWork W2263302838 @default.
• W3117245434 hasRelatedWork W2367214106 @default.
• W3117245434 hasRelatedWork W2379065761 @default.
• W3117245434 hasRelatedWork W2549006548 @default.

• next