Matches in SemOpenAlex for { <https://semopenalex.org/work/W2808399504> ?p ?o ?g. }
Showing items 1 to 94 of
94
with 100 items per page.
- W2808399504 abstract "There is growing evidence that converting targets to soft targets in supervised learning can provide considerable gains in performance. Much of this work has considered classification, converting hard zero-one values to soft labels---such as by adding label noise, incorporating label ambiguity or using distillation. In parallel, there is some evidence from a regression setting in reinforcement learning that learning distributions can improve performance. In this work, we investigate the reasons for this improvement, in a regression setting. We introduce a novel distributional regression loss, and similarly find it significantly improves prediction accuracy. We investigate several common hypotheses, around reducing overfitting and improved representations. We instead find evidence for an alternative hypothesis: this loss is easier to optimize, with better behaved gradients, resulting in improved generalization. We provide theoretical support for this alternative hypothesis, by characterizing the norm of the gradients of this loss." @default.
- W2808399504 created "2018-06-21" @default.
- W2808399504 creator A5025265456 @default.
- W2808399504 creator A5031606152 @default.
- W2808399504 date "2018-06-12" @default.
- W2808399504 modified "2023-09-27" @default.
- W2808399504 title "Improving Regression Performance with Distributional Losses." @default.
- W2808399504 cites W144691381 @default.
- W2808399504 cites W1522301498 @default.
- W2808399504 cites W1556219185 @default.
- W2808399504 cites W1579853615 @default.
- W2808399504 cites W1811750039 @default.
- W2808399504 cites W1821462560 @default.
- W2808399504 cites W2006398000 @default.
- W2808399504 cites W2095705004 @default.
- W2808399504 cites W2101234009 @default.
- W2808399504 cites W2135354436 @default.
- W2808399504 cites W2183341477 @default.
- W2808399504 cites W2300770211 @default.
- W2808399504 cites W2508728158 @default.
- W2808399504 cites W2578804018 @default.
- W2808399504 cites W2604156156 @default.
- W2808399504 cites W2739473244 @default.
- W2808399504 cites W2950300355 @default.
- W2808399504 cites W2950950272 @default.
- W2808399504 cites W2952881492 @default.
- W2808399504 cites W2953318193 @default.
- W2808399504 cites W2963238274 @default.
- W2808399504 cites W2963242170 @default.
- W2808399504 cites W3215037115 @default.
- W2808399504 hasPublicationYear "2018" @default.
- W2808399504 type Work @default.
- W2808399504 sameAs 2808399504 @default.
- W2808399504 citedByCount "6" @default.
- W2808399504 countsByYear W28083995042018 @default.
- W2808399504 countsByYear W28083995042019 @default.
- W2808399504 countsByYear W28083995042020 @default.
- W2808399504 countsByYear W28083995042021 @default.
- W2808399504 crossrefType "posted-content" @default.
- W2808399504 hasAuthorship W2808399504A5025265456 @default.
- W2808399504 hasAuthorship W2808399504A5031606152 @default.
- W2808399504 hasConcept C105795698 @default.
- W2808399504 hasConcept C119857082 @default.
- W2808399504 hasConcept C134306372 @default.
- W2808399504 hasConcept C154945302 @default.
- W2808399504 hasConcept C177148314 @default.
- W2808399504 hasConcept C199360897 @default.
- W2808399504 hasConcept C22019652 @default.
- W2808399504 hasConcept C2780522230 @default.
- W2808399504 hasConcept C33923547 @default.
- W2808399504 hasConcept C41008148 @default.
- W2808399504 hasConcept C50644808 @default.
- W2808399504 hasConcept C83546350 @default.
- W2808399504 hasConcept C97541855 @default.
- W2808399504 hasConceptScore W2808399504C105795698 @default.
- W2808399504 hasConceptScore W2808399504C119857082 @default.
- W2808399504 hasConceptScore W2808399504C134306372 @default.
- W2808399504 hasConceptScore W2808399504C154945302 @default.
- W2808399504 hasConceptScore W2808399504C177148314 @default.
- W2808399504 hasConceptScore W2808399504C199360897 @default.
- W2808399504 hasConceptScore W2808399504C22019652 @default.
- W2808399504 hasConceptScore W2808399504C2780522230 @default.
- W2808399504 hasConceptScore W2808399504C33923547 @default.
- W2808399504 hasConceptScore W2808399504C41008148 @default.
- W2808399504 hasConceptScore W2808399504C50644808 @default.
- W2808399504 hasConceptScore W2808399504C83546350 @default.
- W2808399504 hasConceptScore W2808399504C97541855 @default.
- W2808399504 hasLocation W28083995041 @default.
- W2808399504 hasOpenAccess W2808399504 @default.
- W2808399504 hasPrimaryLocation W28083995041 @default.
- W2808399504 hasRelatedWork W1966992081 @default.
- W2808399504 hasRelatedWork W2151989798 @default.
- W2808399504 hasRelatedWork W2187827799 @default.
- W2808399504 hasRelatedWork W2736081685 @default.
- W2808399504 hasRelatedWork W2736273880 @default.
- W2808399504 hasRelatedWork W2774700420 @default.
- W2808399504 hasRelatedWork W2899682268 @default.
- W2808399504 hasRelatedWork W2963612511 @default.
- W2808399504 hasRelatedWork W3014083813 @default.
- W2808399504 hasRelatedWork W3015036596 @default.
- W2808399504 hasRelatedWork W3035409869 @default.
- W2808399504 hasRelatedWork W3094203027 @default.
- W2808399504 hasRelatedWork W3120211995 @default.
- W2808399504 hasRelatedWork W3135338544 @default.
- W2808399504 hasRelatedWork W3135678629 @default.
- W2808399504 hasRelatedWork W3158223111 @default.
- W2808399504 hasRelatedWork W3173031374 @default.
- W2808399504 hasRelatedWork W3173704566 @default.
- W2808399504 hasRelatedWork W3191940631 @default.
- W2808399504 hasRelatedWork W3208001237 @default.
- W2808399504 isParatext "false" @default.
- W2808399504 isRetracted "false" @default.
- W2808399504 magId "2808399504" @default.
- W2808399504 workType "article" @default.