Matches in SemOpenAlex for { <https://semopenalex.org/work/W2945236295> ?p ?o ?g. }
Showing items 1 to 83 of
83
with 100 items per page.
- W2945236295 abstract "Convolutional neural networks (CNNs) have achieved remarkable performance in various fields, particularly in the domain of computer vision. However, why this architecture works well remains to be a mystery. In this work we move a small step toward understanding the success of CNNs by investigating the learning dynamics of a two-layer nonlinear convolutional neural network over some specific data distributions. Rather than the typical Gaussian assumption for input data distribution, we consider a more realistic setting that each data point (e.g. image) contains a specific pattern determining its class label. Within this setting, we both theoretically and empirically show that some convolutional filters will learn the key patterns in data and the norm of these filters will dominate during the training process with stochastic gradient descent. And with any high probability, when the number of iterations is sufficiently large, the CNN model could obtain 100% accuracy over the considered data distributions. Our experiments demonstrate that for practical image classification tasks our findings still hold to some extent." @default.
- W2945236295 created "2019-05-29" @default.
- W2945236295 creator A5006845236 @default.
- W2945236295 creator A5021441027 @default.
- W2945236295 creator A5045305860 @default.
- W2945236295 date "2019-05-24" @default.
- W2945236295 modified "2023-10-16" @default.
- W2945236295 title "On the Learning Dynamics of Two-layer Nonlinear Convolutional Neural Networks" @default.
- W2945236295 cites W1849277567 @default.
- W2945236295 cites W2257979135 @default.
- W2945236295 cites W2591714514 @default.
- W2945236295 cites W2768041186 @default.
- W2945236295 cites W2786689480 @default.
- W2945236295 cites W2886067286 @default.
- W2945236295 cites W2894604724 @default.
- W2945236295 cites W2898422183 @default.
- W2945236295 cites W2899748887 @default.
- W2945236295 cites W2900959181 @default.
- W2945236295 cites W2911867426 @default.
- W2945236295 cites W2919115771 @default.
- W2945236295 cites W2962698540 @default.
- W2945236295 cites W2963100491 @default.
- W2945236295 cites W2963672337 @default.
- W2945236295 cites W2963744427 @default.
- W2945236295 cites W2963826371 @default.
- W2945236295 cites W2963970792 @default.
- W2945236295 cites W937846259 @default.
- W2945236295 doi "https://doi.org/10.48550/arxiv.1905.10157" @default.
- W2945236295 hasPublicationYear "2019" @default.
- W2945236295 type Work @default.
- W2945236295 sameAs 2945236295 @default.
- W2945236295 citedByCount "1" @default.
- W2945236295 countsByYear W29452362952020 @default.
- W2945236295 crossrefType "posted-content" @default.
- W2945236295 hasAuthorship W2945236295A5006845236 @default.
- W2945236295 hasAuthorship W2945236295A5021441027 @default.
- W2945236295 hasAuthorship W2945236295A5045305860 @default.
- W2945236295 hasBestOaLocation W29452362951 @default.
- W2945236295 hasConcept C11413529 @default.
- W2945236295 hasConcept C119857082 @default.
- W2945236295 hasConcept C121332964 @default.
- W2945236295 hasConcept C153180895 @default.
- W2945236295 hasConcept C153258448 @default.
- W2945236295 hasConcept C154945302 @default.
- W2945236295 hasConcept C158622935 @default.
- W2945236295 hasConcept C163716315 @default.
- W2945236295 hasConcept C206688291 @default.
- W2945236295 hasConcept C2777212361 @default.
- W2945236295 hasConcept C41008148 @default.
- W2945236295 hasConcept C50644808 @default.
- W2945236295 hasConcept C62520636 @default.
- W2945236295 hasConcept C81363708 @default.
- W2945236295 hasConceptScore W2945236295C11413529 @default.
- W2945236295 hasConceptScore W2945236295C119857082 @default.
- W2945236295 hasConceptScore W2945236295C121332964 @default.
- W2945236295 hasConceptScore W2945236295C153180895 @default.
- W2945236295 hasConceptScore W2945236295C153258448 @default.
- W2945236295 hasConceptScore W2945236295C154945302 @default.
- W2945236295 hasConceptScore W2945236295C158622935 @default.
- W2945236295 hasConceptScore W2945236295C163716315 @default.
- W2945236295 hasConceptScore W2945236295C206688291 @default.
- W2945236295 hasConceptScore W2945236295C2777212361 @default.
- W2945236295 hasConceptScore W2945236295C41008148 @default.
- W2945236295 hasConceptScore W2945236295C50644808 @default.
- W2945236295 hasConceptScore W2945236295C62520636 @default.
- W2945236295 hasConceptScore W2945236295C81363708 @default.
- W2945236295 hasLocation W29452362951 @default.
- W2945236295 hasOpenAccess W2945236295 @default.
- W2945236295 hasPrimaryLocation W29452362951 @default.
- W2945236295 hasRelatedWork W2175746458 @default.
- W2945236295 hasRelatedWork W2732542196 @default.
- W2945236295 hasRelatedWork W2738221750 @default.
- W2945236295 hasRelatedWork W2760085659 @default.
- W2945236295 hasRelatedWork W2883200793 @default.
- W2945236295 hasRelatedWork W3012978760 @default.
- W2945236295 hasRelatedWork W3027997911 @default.
- W2945236295 hasRelatedWork W3038075430 @default.
- W2945236295 hasRelatedWork W3093612317 @default.
- W2945236295 hasRelatedWork W4287776258 @default.
- W2945236295 isParatext "false" @default.
- W2945236295 isRetracted "false" @default.
- W2945236295 magId "2945236295" @default.
- W2945236295 workType "article" @default.