Matches in SemOpenAlex for { <https://semopenalex.org/work/W3046756688> ?p ?o ?g. }
Showing items 1 to 79 of
79
with 100 items per page.
- W3046756688 abstract "The classical Universal Approximation Theorem holds for neural networks of arbitrary width and bounded depth. Here we consider the natural `dual' scenario for networks of bounded width and arbitrary depth. Precisely, let $n$ be the number of inputs neurons, $m$ be the number of output neurons, and let $rho$ be any nonaffine continuous function, with a continuous nonzero derivative at some point. Then we show that the class of neural networks of arbitrary depth, width $n + m + 2$, and activation function $rho$, is dense in $C(K; mathbb{R}^m)$ for $K subseteq mathbb{R}^n$ with $K$ compact. This covers every activation function possible to use in practice, and also includes polynomial activation functions, which is unlike the classical version of the theorem, and provides a qualitative difference between deep narrow networks and shallow wide networks. We then consider several extensions of this result. In particular we consider nowhere differentiable activation functions, density in noncompact domains with respect to the $L^p$-norm, and how the width may be reduced to just $n + m + 1$ for `most' activation functions." @default.
- W3046756688 created "2020-08-07" @default.
- W3046756688 creator A5001671573 @default.
- W3046756688 creator A5033716148 @default.
- W3046756688 date "2019-09-25" @default.
- W3046756688 modified "2023-09-27" @default.
- W3046756688 title "Universal approximation with deep narrow networks" @default.
- W3046756688 hasPublicationYear "2019" @default.
- W3046756688 type Work @default.
- W3046756688 sameAs 3046756688 @default.
- W3046756688 citedByCount "0" @default.
- W3046756688 crossrefType "journal-article" @default.
- W3046756688 hasAuthorship W3046756688A5001671573 @default.
- W3046756688 hasAuthorship W3046756688A5033716148 @default.
- W3046756688 hasConcept C114614502 @default.
- W3046756688 hasConcept C118615104 @default.
- W3046756688 hasConcept C119857082 @default.
- W3046756688 hasConcept C134306372 @default.
- W3046756688 hasConcept C14036430 @default.
- W3046756688 hasConcept C17744445 @default.
- W3046756688 hasConcept C191795146 @default.
- W3046756688 hasConcept C199539241 @default.
- W3046756688 hasConcept C202444582 @default.
- W3046756688 hasConcept C202615002 @default.
- W3046756688 hasConcept C33923547 @default.
- W3046756688 hasConcept C34388435 @default.
- W3046756688 hasConcept C38365724 @default.
- W3046756688 hasConcept C41008148 @default.
- W3046756688 hasConcept C50644808 @default.
- W3046756688 hasConcept C78458016 @default.
- W3046756688 hasConcept C86803240 @default.
- W3046756688 hasConcept C90119067 @default.
- W3046756688 hasConceptScore W3046756688C114614502 @default.
- W3046756688 hasConceptScore W3046756688C118615104 @default.
- W3046756688 hasConceptScore W3046756688C119857082 @default.
- W3046756688 hasConceptScore W3046756688C134306372 @default.
- W3046756688 hasConceptScore W3046756688C14036430 @default.
- W3046756688 hasConceptScore W3046756688C17744445 @default.
- W3046756688 hasConceptScore W3046756688C191795146 @default.
- W3046756688 hasConceptScore W3046756688C199539241 @default.
- W3046756688 hasConceptScore W3046756688C202444582 @default.
- W3046756688 hasConceptScore W3046756688C202615002 @default.
- W3046756688 hasConceptScore W3046756688C33923547 @default.
- W3046756688 hasConceptScore W3046756688C34388435 @default.
- W3046756688 hasConceptScore W3046756688C38365724 @default.
- W3046756688 hasConceptScore W3046756688C41008148 @default.
- W3046756688 hasConceptScore W3046756688C50644808 @default.
- W3046756688 hasConceptScore W3046756688C78458016 @default.
- W3046756688 hasConceptScore W3046756688C86803240 @default.
- W3046756688 hasConceptScore W3046756688C90119067 @default.
- W3046756688 hasIssue "2020" @default.
- W3046756688 hasLocation W30467566881 @default.
- W3046756688 hasOpenAccess W3046756688 @default.
- W3046756688 hasPrimaryLocation W30467566881 @default.
- W3046756688 hasRelatedWork W1924327771 @default.
- W3046756688 hasRelatedWork W1993364643 @default.
- W3046756688 hasRelatedWork W1998452661 @default.
- W3046756688 hasRelatedWork W2000585002 @default.
- W3046756688 hasRelatedWork W2080480532 @default.
- W3046756688 hasRelatedWork W2244057755 @default.
- W3046756688 hasRelatedWork W2530549011 @default.
- W3046756688 hasRelatedWork W2897154110 @default.
- W3046756688 hasRelatedWork W2913291429 @default.
- W3046756688 hasRelatedWork W2942397257 @default.
- W3046756688 hasRelatedWork W2946077927 @default.
- W3046756688 hasRelatedWork W2949907429 @default.
- W3046756688 hasRelatedWork W2956571163 @default.
- W3046756688 hasRelatedWork W2963933465 @default.
- W3046756688 hasRelatedWork W3099934582 @default.
- W3046756688 hasRelatedWork W3111213259 @default.
- W3046756688 hasRelatedWork W3133838324 @default.
- W3046756688 hasRelatedWork W3174418817 @default.
- W3046756688 hasRelatedWork W3186073754 @default.
- W3046756688 hasRelatedWork W3212105521 @default.
- W3046756688 hasVolume "125" @default.
- W3046756688 isParatext "false" @default.
- W3046756688 isRetracted "false" @default.
- W3046756688 magId "3046756688" @default.
- W3046756688 workType "article" @default.