Matches in SemOpenAlex for { <https://semopenalex.org/work/W2944856801> ?p ?o ?g. }
- W2944856801 abstract "To help understand the underlying mechanisms of neural networks (NNs), several groups have, in recent years, studied the number of linear regions $ell$ of piecewise linear functions generated by deep neural networks (DNN). In particular, they showed that $ell$ can grow exponentially with the number of network parameters $p$, a property often used to explain the advantages of DNNs over shallow NNs in approximating complicated functions. Nonetheless, a simple dimension argument shows that DNNs cannot generate all piecewise linear functions with $ell$ linear regions as soon as $ell > p$. It is thus natural to seek to characterize specific families of functions with $ell$ linear regions that can be constructed by DNNs. Iterated Function Systems (IFS) generate sequences of piecewise linear functions $F_k$ with a number of linear regions exponential in $k$. We show that, under mild assumptions, $F_k$ can be generated by a NN using only $mathcal{O}(k)$ parameters. IFS are used extensively to generate, at low computational cost, natural-looking landscape textures in artificial images. They have also been proposed for compression of natural images, albeit with less commercial success. The surprisingly good performance of this fractal-based compression suggests that our visual system may lock in, to some extent, on self-similarities in images. The combination of this phenomenon with the capacity, demonstrated here, of DNNs to efficiently approximate IFS may contribute to the success of DNNs, particularly striking for image processing tasks, as well as suggest new algorithms for representing self similarities in images based on the DNN mechanism." @default.
- W2944856801 created "2019-05-29" @default.
- W2944856801 creator A5074615861 @default.
- W2944856801 creator A5078787909 @default.
- W2944856801 creator A5086417185 @default.
- W2944856801 date "2019-05-27" @default.
- W2944856801 modified "2023-09-27" @default.
- W2944856801 title "Expression of Fractals Through Neural Network Functions." @default.
- W2944856801 cites W1498909268 @default.
- W2944856801 cites W1514643409 @default.
- W2944856801 cites W1972567154 @default.
- W2944856801 cites W2032121576 @default.
- W2944856801 cites W2103496339 @default.
- W2944856801 cites W2105732809 @default.
- W2944856801 cites W2108069432 @default.
- W2944856801 cites W2117130368 @default.
- W2944856801 cites W2123442489 @default.
- W2944856801 cites W2149961130 @default.
- W2944856801 cites W2161388792 @default.
- W2944856801 cites W2163605009 @default.
- W2944856801 cites W2167191464 @default.
- W2944856801 cites W2339378878 @default.
- W2944856801 cites W2420631737 @default.
- W2944856801 cites W2528305538 @default.
- W2944856801 cites W2534320940 @default.
- W2944856801 cites W2550848904 @default.
- W2944856801 cites W2560609797 @default.
- W2944856801 cites W2767609493 @default.
- W2944856801 cites W2910155483 @default.
- W2944856801 cites W2944394121 @default.
- W2944856801 cites W2951548721 @default.
- W2944856801 cites W2962742960 @default.
- W2944856801 cites W2962782102 @default.
- W2944856801 cites W2962845550 @default.
- W2944856801 cites W2971101288 @default.
- W2944856801 cites W3146803896 @default.
- W2944856801 hasPublicationYear "2019" @default.
- W2944856801 type Work @default.
- W2944856801 sameAs 2944856801 @default.
- W2944856801 citedByCount "1" @default.
- W2944856801 countsByYear W29448568012020 @default.
- W2944856801 crossrefType "posted-content" @default.
- W2944856801 hasAuthorship W2944856801A5074615861 @default.
- W2944856801 hasAuthorship W2944856801A5078787909 @default.
- W2944856801 hasAuthorship W2944856801A5086417185 @default.
- W2944856801 hasConcept C111472728 @default.
- W2944856801 hasConcept C11413529 @default.
- W2944856801 hasConcept C118615104 @default.
- W2944856801 hasConcept C134306372 @default.
- W2944856801 hasConcept C138885662 @default.
- W2944856801 hasConcept C14036430 @default.
- W2944856801 hasConcept C140479938 @default.
- W2944856801 hasConcept C154945302 @default.
- W2944856801 hasConcept C168460219 @default.
- W2944856801 hasConcept C17095337 @default.
- W2944856801 hasConcept C2780586882 @default.
- W2944856801 hasConcept C33923547 @default.
- W2944856801 hasConcept C40636538 @default.
- W2944856801 hasConcept C41008148 @default.
- W2944856801 hasConcept C50644808 @default.
- W2944856801 hasConcept C78458016 @default.
- W2944856801 hasConcept C86803240 @default.
- W2944856801 hasConceptScore W2944856801C111472728 @default.
- W2944856801 hasConceptScore W2944856801C11413529 @default.
- W2944856801 hasConceptScore W2944856801C118615104 @default.
- W2944856801 hasConceptScore W2944856801C134306372 @default.
- W2944856801 hasConceptScore W2944856801C138885662 @default.
- W2944856801 hasConceptScore W2944856801C14036430 @default.
- W2944856801 hasConceptScore W2944856801C140479938 @default.
- W2944856801 hasConceptScore W2944856801C154945302 @default.
- W2944856801 hasConceptScore W2944856801C168460219 @default.
- W2944856801 hasConceptScore W2944856801C17095337 @default.
- W2944856801 hasConceptScore W2944856801C2780586882 @default.
- W2944856801 hasConceptScore W2944856801C33923547 @default.
- W2944856801 hasConceptScore W2944856801C40636538 @default.
- W2944856801 hasConceptScore W2944856801C41008148 @default.
- W2944856801 hasConceptScore W2944856801C50644808 @default.
- W2944856801 hasConceptScore W2944856801C78458016 @default.
- W2944856801 hasConceptScore W2944856801C86803240 @default.
- W2944856801 hasLocation W29448568011 @default.
- W2944856801 hasOpenAccess W2944856801 @default.
- W2944856801 hasPrimaryLocation W29448568011 @default.
- W2944856801 hasRelatedWork W1565765145 @default.
- W2944856801 hasRelatedWork W2317787243 @default.
- W2944856801 hasRelatedWork W2552459341 @default.
- W2944856801 hasRelatedWork W2552729449 @default.
- W2944856801 hasRelatedWork W2724501717 @default.
- W2944856801 hasRelatedWork W2739309726 @default.
- W2944856801 hasRelatedWork W2766808838 @default.
- W2944856801 hasRelatedWork W2809468328 @default.
- W2944856801 hasRelatedWork W2811074806 @default.
- W2944856801 hasRelatedWork W2811254153 @default.
- W2944856801 hasRelatedWork W2884522222 @default.
- W2944856801 hasRelatedWork W2949084874 @default.
- W2944856801 hasRelatedWork W2951603627 @default.
- W2944856801 hasRelatedWork W2970243504 @default.
- W2944856801 hasRelatedWork W3034848862 @default.
- W2944856801 hasRelatedWork W3037559325 @default.
- W2944856801 hasRelatedWork W3094965745 @default.
- W2944856801 hasRelatedWork W3099937123 @default.