Matches in SemOpenAlex for { <https://semopenalex.org/work/W2767609493> ?p ?o ?g. }
- W2767609493 abstract "We investigate the complexity of deep neural networks (DNN) that represent piecewise linear (PWL) functions. In particular, we study the number of linear regions, i.e. pieces, that a PWL function represented by a DNN can attain, both theoretically and empirically. We present (i) tighter upper and lower bounds for the maximum number of linear regions on rectifier networks, which are exact for inputs of dimension one; (ii) a first upper bound for multi-layer maxout networks; and (iii) a first method to perform exact enumeration or counting of the number of regions by modeling the DNN with a mixed-integer linear formulation. These bounds come from leveraging the dimension of the space defining each linear region. The results also indicate that a deep rectifier network can only have more linear regions than every shallow counterpart with same number of neurons if that number exceeds the dimension of the input." @default.
- W2767609493 created "2017-11-17" @default.
- W2767609493 creator A5017504613 @default.
- W2767609493 creator A5021227744 @default.
- W2767609493 creator A5060350100 @default.
- W2767609493 date "2017-11-06" @default.
- W2767609493 modified "2023-09-27" @default.
- W2767609493 title "Bounding and Counting Linear Regions of Deep Neural Networks" @default.
- W2767609493 cites W1484418787 @default.
- W2767609493 cites W1542886316 @default.
- W2767609493 cites W1885159080 @default.
- W2767609493 cites W1970193303 @default.
- W2767609493 cites W1981530182 @default.
- W2767609493 cites W2034276439 @default.
- W2767609493 cites W2039415399 @default.
- W2767609493 cites W2080263450 @default.
- W2767609493 cites W2089947415 @default.
- W2767609493 cites W2097117768 @default.
- W2767609493 cites W2103496339 @default.
- W2767609493 cites W2112796928 @default.
- W2767609493 cites W2141473882 @default.
- W2767609493 cites W2156822338 @default.
- W2767609493 cites W2161388792 @default.
- W2767609493 cites W2184045248 @default.
- W2767609493 cites W2194775991 @default.
- W2767609493 cites W2288433034 @default.
- W2767609493 cites W2962742960 @default.
- W2767609493 cites W2963982496 @default.
- W2767609493 cites W3037950864 @default.
- W2767609493 cites W3044803474 @default.
- W2767609493 hasPublicationYear "2017" @default.
- W2767609493 type Work @default.
- W2767609493 sameAs 2767609493 @default.
- W2767609493 citedByCount "23" @default.
- W2767609493 countsByYear W27676094932016 @default.
- W2767609493 countsByYear W27676094932017 @default.
- W2767609493 countsByYear W27676094932018 @default.
- W2767609493 countsByYear W27676094932019 @default.
- W2767609493 countsByYear W27676094932020 @default.
- W2767609493 countsByYear W27676094932021 @default.
- W2767609493 crossrefType "posted-content" @default.
- W2767609493 hasAuthorship W2767609493A5017504613 @default.
- W2767609493 hasAuthorship W2767609493A5021227744 @default.
- W2767609493 hasAuthorship W2767609493A5060350100 @default.
- W2767609493 hasConcept C11413529 @default.
- W2767609493 hasConcept C114614502 @default.
- W2767609493 hasConcept C119857082 @default.
- W2767609493 hasConcept C134306372 @default.
- W2767609493 hasConcept C14036430 @default.
- W2767609493 hasConcept C154945302 @default.
- W2767609493 hasConcept C156340839 @default.
- W2767609493 hasConcept C163175372 @default.
- W2767609493 hasConcept C17095337 @default.
- W2767609493 hasConcept C175202392 @default.
- W2767609493 hasConcept C176370821 @default.
- W2767609493 hasConcept C199360897 @default.
- W2767609493 hasConcept C33676613 @default.
- W2767609493 hasConcept C33923547 @default.
- W2767609493 hasConcept C41008148 @default.
- W2767609493 hasConcept C41045048 @default.
- W2767609493 hasConcept C50100734 @default.
- W2767609493 hasConcept C50644808 @default.
- W2767609493 hasConcept C63584917 @default.
- W2767609493 hasConcept C77553402 @default.
- W2767609493 hasConcept C78458016 @default.
- W2767609493 hasConcept C86582703 @default.
- W2767609493 hasConcept C86803240 @default.
- W2767609493 hasConcept C97137487 @default.
- W2767609493 hasConceptScore W2767609493C11413529 @default.
- W2767609493 hasConceptScore W2767609493C114614502 @default.
- W2767609493 hasConceptScore W2767609493C119857082 @default.
- W2767609493 hasConceptScore W2767609493C134306372 @default.
- W2767609493 hasConceptScore W2767609493C14036430 @default.
- W2767609493 hasConceptScore W2767609493C154945302 @default.
- W2767609493 hasConceptScore W2767609493C156340839 @default.
- W2767609493 hasConceptScore W2767609493C163175372 @default.
- W2767609493 hasConceptScore W2767609493C17095337 @default.
- W2767609493 hasConceptScore W2767609493C175202392 @default.
- W2767609493 hasConceptScore W2767609493C176370821 @default.
- W2767609493 hasConceptScore W2767609493C199360897 @default.
- W2767609493 hasConceptScore W2767609493C33676613 @default.
- W2767609493 hasConceptScore W2767609493C33923547 @default.
- W2767609493 hasConceptScore W2767609493C41008148 @default.
- W2767609493 hasConceptScore W2767609493C41045048 @default.
- W2767609493 hasConceptScore W2767609493C50100734 @default.
- W2767609493 hasConceptScore W2767609493C50644808 @default.
- W2767609493 hasConceptScore W2767609493C63584917 @default.
- W2767609493 hasConceptScore W2767609493C77553402 @default.
- W2767609493 hasConceptScore W2767609493C78458016 @default.
- W2767609493 hasConceptScore W2767609493C86582703 @default.
- W2767609493 hasConceptScore W2767609493C86803240 @default.
- W2767609493 hasConceptScore W2767609493C97137487 @default.
- W2767609493 hasLocation W27676094931 @default.
- W2767609493 hasOpenAccess W2767609493 @default.
- W2767609493 hasPrimaryLocation W27676094931 @default.
- W2767609493 hasRelatedWork W1981530182 @default.
- W2767609493 hasRelatedWork W2134974549 @default.
- W2767609493 hasRelatedWork W2161388792 @default.
- W2767609493 hasRelatedWork W2208555118 @default.
- W2767609493 hasRelatedWork W2281746805 @default.