Matches in SemOpenAlex for { <https://semopenalex.org/work/W3035253990> ?p ?o ?g. }
Showing items 1 to 73 of
73
with 100 items per page.
- W3035253990 abstract "Causal modeling has been recognized as a potential solution to many challenging problems in machine learning (ML). Here, we propose a counterfactual approach to remove/reduce the influence of confounders from the predictions generated a deep neural network (DNN). Rather than attempting to prevent DNNs from directly learning the confounding signal, we propose a counterfactual approach to remove confounding from the feature representations learned by DNNs in anticausal prediction tasks. By training an accurate DNN using softmax activation at the classification layer, and then adopting the representation learned by the last layer prior to the output layer as our features, we have that, by construction, the learned features will fit well a logistic regression model, and will be linearly associated with the labels. Then, in order to generate classifiers that are free from the influence of the observed confounders we: (i) use linear models to regress each learned feature on the labels and on the confounders and estimate the respective regression coefficients and model residuals; (ii) generate new counterfactual features by adding back to the estimated residuals to a linear predictor which no longer includes the confounder variables; and (iii) train and evaluate a logistic classifier using the counterfactual features as inputs. We validate the proposed methodology using colored versions of the MNIST and fashion-MNIST datasets, and show how the approach can effectively combat confounding and improve generalization in the context of dataset shift. Comparison against a variation of the SMOTE cite{chawla2002} approach showed that the causality-aware approach compared favorably against SMOTE balancing in our experiments. Finally, we also describe how to use conditional independence tests to evaluate if the counterfactual approach has effectively removed the confounder signals from the predictions." @default.
- W3035253990 created "2020-06-19" @default.
- W3035253990 creator A5029959684 @default.
- W3035253990 date "2020-04-20" @default.
- W3035253990 modified "2023-09-27" @default.
- W3035253990 title "Causality-aware counterfactual confounding adjustment for feature representations learned by deep models: with an application to image classification tasks" @default.
- W3035253990 hasPublicationYear "2020" @default.
- W3035253990 type Work @default.
- W3035253990 sameAs 3035253990 @default.
- W3035253990 citedByCount "0" @default.
- W3035253990 crossrefType "posted-content" @default.
- W3035253990 hasAuthorship W3035253990A5029959684 @default.
- W3035253990 hasConcept C105795698 @default.
- W3035253990 hasConcept C108650721 @default.
- W3035253990 hasConcept C119857082 @default.
- W3035253990 hasConcept C138885662 @default.
- W3035253990 hasConcept C151956035 @default.
- W3035253990 hasConcept C153180895 @default.
- W3035253990 hasConcept C154945302 @default.
- W3035253990 hasConcept C15744967 @default.
- W3035253990 hasConcept C188441871 @default.
- W3035253990 hasConcept C190502265 @default.
- W3035253990 hasConcept C2776401178 @default.
- W3035253990 hasConcept C33923547 @default.
- W3035253990 hasConcept C41008148 @default.
- W3035253990 hasConcept C41895202 @default.
- W3035253990 hasConcept C50644808 @default.
- W3035253990 hasConcept C77350462 @default.
- W3035253990 hasConcept C77805123 @default.
- W3035253990 hasConceptScore W3035253990C105795698 @default.
- W3035253990 hasConceptScore W3035253990C108650721 @default.
- W3035253990 hasConceptScore W3035253990C119857082 @default.
- W3035253990 hasConceptScore W3035253990C138885662 @default.
- W3035253990 hasConceptScore W3035253990C151956035 @default.
- W3035253990 hasConceptScore W3035253990C153180895 @default.
- W3035253990 hasConceptScore W3035253990C154945302 @default.
- W3035253990 hasConceptScore W3035253990C15744967 @default.
- W3035253990 hasConceptScore W3035253990C188441871 @default.
- W3035253990 hasConceptScore W3035253990C190502265 @default.
- W3035253990 hasConceptScore W3035253990C2776401178 @default.
- W3035253990 hasConceptScore W3035253990C33923547 @default.
- W3035253990 hasConceptScore W3035253990C41008148 @default.
- W3035253990 hasConceptScore W3035253990C41895202 @default.
- W3035253990 hasConceptScore W3035253990C50644808 @default.
- W3035253990 hasConceptScore W3035253990C77350462 @default.
- W3035253990 hasConceptScore W3035253990C77805123 @default.
- W3035253990 hasLocation W30352539901 @default.
- W3035253990 hasOpenAccess W3035253990 @default.
- W3035253990 hasPrimaryLocation W30352539901 @default.
- W3035253990 hasRelatedWork W1586978959 @default.
- W3035253990 hasRelatedWork W2104552956 @default.
- W3035253990 hasRelatedWork W2111361857 @default.
- W3035253990 hasRelatedWork W2111947236 @default.
- W3035253990 hasRelatedWork W2279253867 @default.
- W3035253990 hasRelatedWork W2527333635 @default.
- W3035253990 hasRelatedWork W2790768383 @default.
- W3035253990 hasRelatedWork W2883670420 @default.
- W3035253990 hasRelatedWork W2937523352 @default.
- W3035253990 hasRelatedWork W2945885813 @default.
- W3035253990 hasRelatedWork W3003896119 @default.
- W3035253990 hasRelatedWork W3083953438 @default.
- W3035253990 hasRelatedWork W3115199521 @default.
- W3035253990 hasRelatedWork W3133924939 @default.
- W3035253990 hasRelatedWork W3194060937 @default.
- W3035253990 hasRelatedWork W3199150760 @default.
- W3035253990 hasRelatedWork W3211589367 @default.
- W3035253990 hasRelatedWork W56995616 @default.
- W3035253990 hasRelatedWork W79574811 @default.
- W3035253990 hasRelatedWork W2963683258 @default.
- W3035253990 isParatext "false" @default.
- W3035253990 isRetracted "false" @default.
- W3035253990 magId "3035253990" @default.
- W3035253990 workType "article" @default.