Matches in SemOpenAlex for { <https://semopenalex.org/work/W3129086097> ?p ?o ?g. }
Showing items 1 to 74 of
74
with 100 items per page.
- W3129086097 abstract "Due to spurious correlations, machine learning systems often fail to generalize to environments whose distributions differ from the ones used at training time. Prior work addressing this, either explicitly or implicitly, attempted to find a data representation that has an invariant causal relationship with the outcome. This is done by leveraging a diverse set of training environments to reduce the effect of spurious features, on top of which an invariant classifier is then built. However, these methods have generalization guarantees only when both data representation and classifiers come from a linear model class. As an alternative, we propose Invariant Causal Representation Learning (ICRL), a learning paradigm that enables out-of-distribution generalization in the nonlinear setting (i.e., nonlinear representations and nonlinear classifiers). It builds upon a practical and general assumption: data representations factorize when conditioning on the outcome and the environment. Based on this, we show identifiability up to a permutation and pointwise transformation. We also prove that all direct causes of the outcome can be fully discovered, which further enables us to obtain generalization guarantees in the nonlinear setting. Extensive experiments on both synthetic and real-world datasets show that our approach significantly outperforms a variety of baseline methods." @default.
- W3129086097 created "2021-02-15" @default.
- W3129086097 creator A5024901763 @default.
- W3129086097 creator A5029396252 @default.
- W3129086097 creator A5034013066 @default.
- W3129086097 creator A5044005697 @default.
- W3129086097 date "2021-05-04" @default.
- W3129086097 modified "2023-09-27" @default.
- W3129086097 title "Invariant Causal Representation Learning" @default.
- W3129086097 hasPublicationYear "2021" @default.
- W3129086097 type Work @default.
- W3129086097 sameAs 3129086097 @default.
- W3129086097 citedByCount "1" @default.
- W3129086097 countsByYear W31290860972021 @default.
- W3129086097 crossrefType "journal-article" @default.
- W3129086097 hasAuthorship W3129086097A5024901763 @default.
- W3129086097 hasAuthorship W3129086097A5029396252 @default.
- W3129086097 hasAuthorship W3129086097A5034013066 @default.
- W3129086097 hasAuthorship W3129086097A5044005697 @default.
- W3129086097 hasConcept C119857082 @default.
- W3129086097 hasConcept C121332964 @default.
- W3129086097 hasConcept C122770356 @default.
- W3129086097 hasConcept C134306372 @default.
- W3129086097 hasConcept C154945302 @default.
- W3129086097 hasConcept C158622935 @default.
- W3129086097 hasConcept C190470478 @default.
- W3129086097 hasConcept C2777984123 @default.
- W3129086097 hasConcept C33923547 @default.
- W3129086097 hasConcept C37914503 @default.
- W3129086097 hasConcept C41008148 @default.
- W3129086097 hasConcept C59404180 @default.
- W3129086097 hasConcept C62520636 @default.
- W3129086097 hasConcept C97256817 @default.
- W3129086097 hasConceptScore W3129086097C119857082 @default.
- W3129086097 hasConceptScore W3129086097C121332964 @default.
- W3129086097 hasConceptScore W3129086097C122770356 @default.
- W3129086097 hasConceptScore W3129086097C134306372 @default.
- W3129086097 hasConceptScore W3129086097C154945302 @default.
- W3129086097 hasConceptScore W3129086097C158622935 @default.
- W3129086097 hasConceptScore W3129086097C190470478 @default.
- W3129086097 hasConceptScore W3129086097C2777984123 @default.
- W3129086097 hasConceptScore W3129086097C33923547 @default.
- W3129086097 hasConceptScore W3129086097C37914503 @default.
- W3129086097 hasConceptScore W3129086097C41008148 @default.
- W3129086097 hasConceptScore W3129086097C59404180 @default.
- W3129086097 hasConceptScore W3129086097C62520636 @default.
- W3129086097 hasConceptScore W3129086097C97256817 @default.
- W3129086097 hasLocation W31290860971 @default.
- W3129086097 hasOpenAccess W3129086097 @default.
- W3129086097 hasPrimaryLocation W31290860971 @default.
- W3129086097 hasRelatedWork W147969812 @default.
- W3129086097 hasRelatedWork W1960700909 @default.
- W3129086097 hasRelatedWork W2751825796 @default.
- W3129086097 hasRelatedWork W2787551658 @default.
- W3129086097 hasRelatedWork W2803766319 @default.
- W3129086097 hasRelatedWork W2943490843 @default.
- W3129086097 hasRelatedWork W2949750828 @default.
- W3129086097 hasRelatedWork W2951736998 @default.
- W3129086097 hasRelatedWork W2962848407 @default.
- W3129086097 hasRelatedWork W2965769472 @default.
- W3129086097 hasRelatedWork W2995823288 @default.
- W3129086097 hasRelatedWork W3006253006 @default.
- W3129086097 hasRelatedWork W3033770268 @default.
- W3129086097 hasRelatedWork W3036815251 @default.
- W3129086097 hasRelatedWork W3037364847 @default.
- W3129086097 hasRelatedWork W3120700377 @default.
- W3129086097 hasRelatedWork W3166265366 @default.
- W3129086097 hasRelatedWork W3204447533 @default.
- W3129086097 hasRelatedWork W3208832035 @default.
- W3129086097 hasRelatedWork W3210469818 @default.
- W3129086097 isParatext "false" @default.
- W3129086097 isRetracted "false" @default.
- W3129086097 magId "3129086097" @default.
- W3129086097 workType "article" @default.