Matches in SemOpenAlex for { <https://semopenalex.org/work/W3168889651> ?p ?o ?g. }
Showing items 1 to 73 of
73
with 100 items per page.
- W3168889651 endingPage "12512" @default.
- W3168889651 startingPage "12501" @default.
- W3168889651 abstract "Many weakly supervised classification methods employ a noise transition matrix to capture the class-conditional label corruption. To estimate the transition matrix from noisy data, existing methods often need to estimate the noisy class-posterior, which could be unreliable due to the overconfidence of neural networks. In this work, we propose a theoretically grounded method that can estimate the noise transition matrix and learn a classifier simultaneously, without relying on the error-prone noisy class-posterior estimation. Concretely, inspired by the characteristics of the stochastic label corruption process, we propose total variation regularization, which encourages the predicted probabilities to be more distinguishable from each other. Under mild assumptions, the proposed method yields a consistent estimator of the transition matrix. We show the effectiveness of the proposed method through experiments on benchmark and real-world datasets." @default.
- W3168889651 created "2021-06-22" @default.
- W3168889651 creator A5007098537 @default.
- W3168889651 creator A5015157301 @default.
- W3168889651 creator A5072744508 @default.
- W3168889651 date "2021-07-18" @default.
- W3168889651 modified "2023-10-03" @default.
- W3168889651 title "Learning Noise Transition Matrix from Only Noisy Labels via Total Variation Regularization" @default.
- W3168889651 hasPublicationYear "2021" @default.
- W3168889651 type Work @default.
- W3168889651 sameAs 3168889651 @default.
- W3168889651 citedByCount "4" @default.
- W3168889651 countsByYear W31688896512020 @default.
- W3168889651 countsByYear W31688896512021 @default.
- W3168889651 crossrefType "proceedings-article" @default.
- W3168889651 hasAuthorship W3168889651A5007098537 @default.
- W3168889651 hasAuthorship W3168889651A5015157301 @default.
- W3168889651 hasAuthorship W3168889651A5072744508 @default.
- W3168889651 hasConcept C105795698 @default.
- W3168889651 hasConcept C11413529 @default.
- W3168889651 hasConcept C115961682 @default.
- W3168889651 hasConcept C119857082 @default.
- W3168889651 hasConcept C153180895 @default.
- W3168889651 hasConcept C154945302 @default.
- W3168889651 hasConcept C185429906 @default.
- W3168889651 hasConcept C2776135515 @default.
- W3168889651 hasConcept C33923547 @default.
- W3168889651 hasConcept C41008148 @default.
- W3168889651 hasConcept C49555168 @default.
- W3168889651 hasConcept C95623464 @default.
- W3168889651 hasConcept C98763669 @default.
- W3168889651 hasConcept C99498987 @default.
- W3168889651 hasConceptScore W3168889651C105795698 @default.
- W3168889651 hasConceptScore W3168889651C11413529 @default.
- W3168889651 hasConceptScore W3168889651C115961682 @default.
- W3168889651 hasConceptScore W3168889651C119857082 @default.
- W3168889651 hasConceptScore W3168889651C153180895 @default.
- W3168889651 hasConceptScore W3168889651C154945302 @default.
- W3168889651 hasConceptScore W3168889651C185429906 @default.
- W3168889651 hasConceptScore W3168889651C2776135515 @default.
- W3168889651 hasConceptScore W3168889651C33923547 @default.
- W3168889651 hasConceptScore W3168889651C41008148 @default.
- W3168889651 hasConceptScore W3168889651C49555168 @default.
- W3168889651 hasConceptScore W3168889651C95623464 @default.
- W3168889651 hasConceptScore W3168889651C98763669 @default.
- W3168889651 hasConceptScore W3168889651C99498987 @default.
- W3168889651 hasOpenAccess W3168889651 @default.
- W3168889651 hasRelatedWork W1146205752 @default.
- W3168889651 hasRelatedWork W1790658179 @default.
- W3168889651 hasRelatedWork W2292972218 @default.
- W3168889651 hasRelatedWork W2546314851 @default.
- W3168889651 hasRelatedWork W2550154328 @default.
- W3168889651 hasRelatedWork W2788101720 @default.
- W3168889651 hasRelatedWork W2803256244 @default.
- W3168889651 hasRelatedWork W2952885232 @default.
- W3168889651 hasRelatedWork W2985780925 @default.
- W3168889651 hasRelatedWork W3100754158 @default.
- W3168889651 hasRelatedWork W3102643535 @default.
- W3168889651 hasRelatedWork W3104030484 @default.
- W3168889651 hasRelatedWork W3117203337 @default.
- W3168889651 hasRelatedWork W3126748415 @default.
- W3168889651 hasRelatedWork W3130472572 @default.
- W3168889651 hasRelatedWork W3157298807 @default.
- W3168889651 hasRelatedWork W3173623546 @default.
- W3168889651 hasRelatedWork W852785498 @default.
- W3168889651 hasRelatedWork W3037079993 @default.
- W3168889651 hasRelatedWork W3122805849 @default.
- W3168889651 isParatext "false" @default.
- W3168889651 isRetracted "false" @default.
- W3168889651 magId "3168889651" @default.
- W3168889651 workType "article" @default.