Matches in SemOpenAlex for { <https://semopenalex.org/work/W4287777070> ?p ?o ?g. }
Showing items 1 to 77 of
77
with 100 items per page.
- W4287777070 abstract "When training an estimator such as a neural network for tasks like image denoising, it is often preferred to train one estimator and apply it to all noise levels. The de facto training protocol to achieve this goal is to train the estimator with noisy samples whose noise levels are uniformly distributed across the range of interest. However, why should we allocate the samples uniformly? Can we have more training samples that are less noisy, and fewer samples that are more noisy? What is the optimal distribution? How do we obtain such a distribution? The goal of this paper is to address this training sample distribution problem from a minimax risk optimization perspective. We derive a dual ascent algorithm to determine the optimal sampling distribution of which the convergence is guaranteed as long as the set of admissible estimators is closed and convex. For estimators with non-convex admissible sets such as deep neural networks, our dual formulation converges to a solution of the convex relaxation. We discuss how the algorithm can be implemented in practice. We evaluate the algorithm on linear estimators and deep networks." @default.
- W4287777070 created "2022-07-26" @default.
- W4287777070 creator A5007938446 @default.
- W4287777070 creator A5075004629 @default.
- W4287777070 date "2020-05-19" @default.
- W4287777070 modified "2023-10-16" @default.
- W4287777070 title "One Size Fits All: Can We Train One Denoiser for All Noise Levels?" @default.
- W4287777070 doi "https://doi.org/10.48550/arxiv.2005.09627" @default.
- W4287777070 hasPublicationYear "2020" @default.
- W4287777070 type Work @default.
- W4287777070 citedByCount "0" @default.
- W4287777070 crossrefType "posted-content" @default.
- W4287777070 hasAuthorship W4287777070A5007938446 @default.
- W4287777070 hasAuthorship W4287777070A5075004629 @default.
- W4287777070 hasBestOaLocation W42877770701 @default.
- W4287777070 hasConcept C105795698 @default.
- W4287777070 hasConcept C112680207 @default.
- W4287777070 hasConcept C11413529 @default.
- W4287777070 hasConcept C115961682 @default.
- W4287777070 hasConcept C126255220 @default.
- W4287777070 hasConcept C149728462 @default.
- W4287777070 hasConcept C154945302 @default.
- W4287777070 hasConcept C157972887 @default.
- W4287777070 hasConcept C159985019 @default.
- W4287777070 hasConcept C162324750 @default.
- W4287777070 hasConcept C185429906 @default.
- W4287777070 hasConcept C185592680 @default.
- W4287777070 hasConcept C192562407 @default.
- W4287777070 hasConcept C198531522 @default.
- W4287777070 hasConcept C204323151 @default.
- W4287777070 hasConcept C2524010 @default.
- W4287777070 hasConcept C2777303404 @default.
- W4287777070 hasConcept C33923547 @default.
- W4287777070 hasConcept C41008148 @default.
- W4287777070 hasConcept C43617362 @default.
- W4287777070 hasConcept C50522688 @default.
- W4287777070 hasConcept C50644808 @default.
- W4287777070 hasConcept C99498987 @default.
- W4287777070 hasConceptScore W4287777070C105795698 @default.
- W4287777070 hasConceptScore W4287777070C112680207 @default.
- W4287777070 hasConceptScore W4287777070C11413529 @default.
- W4287777070 hasConceptScore W4287777070C115961682 @default.
- W4287777070 hasConceptScore W4287777070C126255220 @default.
- W4287777070 hasConceptScore W4287777070C149728462 @default.
- W4287777070 hasConceptScore W4287777070C154945302 @default.
- W4287777070 hasConceptScore W4287777070C157972887 @default.
- W4287777070 hasConceptScore W4287777070C159985019 @default.
- W4287777070 hasConceptScore W4287777070C162324750 @default.
- W4287777070 hasConceptScore W4287777070C185429906 @default.
- W4287777070 hasConceptScore W4287777070C185592680 @default.
- W4287777070 hasConceptScore W4287777070C192562407 @default.
- W4287777070 hasConceptScore W4287777070C198531522 @default.
- W4287777070 hasConceptScore W4287777070C204323151 @default.
- W4287777070 hasConceptScore W4287777070C2524010 @default.
- W4287777070 hasConceptScore W4287777070C2777303404 @default.
- W4287777070 hasConceptScore W4287777070C33923547 @default.
- W4287777070 hasConceptScore W4287777070C41008148 @default.
- W4287777070 hasConceptScore W4287777070C43617362 @default.
- W4287777070 hasConceptScore W4287777070C50522688 @default.
- W4287777070 hasConceptScore W4287777070C50644808 @default.
- W4287777070 hasConceptScore W4287777070C99498987 @default.
- W4287777070 hasLocation W42877770701 @default.
- W4287777070 hasOpenAccess W4287777070 @default.
- W4287777070 hasPrimaryLocation W42877770701 @default.
- W4287777070 hasRelatedWork W1591157878 @default.
- W4287777070 hasRelatedWork W1630840116 @default.
- W4287777070 hasRelatedWork W1973573083 @default.
- W4287777070 hasRelatedWork W2022260956 @default.
- W4287777070 hasRelatedWork W2071095524 @default.
- W4287777070 hasRelatedWork W2227983344 @default.
- W4287777070 hasRelatedWork W2356755074 @default.
- W4287777070 hasRelatedWork W2620834727 @default.
- W4287777070 hasRelatedWork W2811008754 @default.
- W4287777070 hasRelatedWork W2486436362 @default.
- W4287777070 isParatext "false" @default.
- W4287777070 isRetracted "false" @default.
- W4287777070 workType "article" @default.