MESSAI.iO

AI summary

50% confidence

The paper explores the use of deep sigmoid neural networks to learn and approximate piecewise smooth functions, demonstrating their effectiveness in approximating complex functions.

Generated by MESSAI extraction pipeline · review against source PDF

Extraction

Reported parameters

90 extracted values

View extracted values

No 3D model is mapped to this paper yet. Parameter ranges above still place reported values on the literature distribution.

Key findings

  • Deep sigmoid neural networks can accurately approximate piecewise smooth functions with a small number of neurons.
  • The proposed method outperforms traditional neural networks in approximating functions with sharp transitions.
  • The network's ability to learn and approximate piecewise smooth functions is attributed to its use of sigmoid activation functions.

Keywords

Sigmoid functionPiecewiseArtificial neural networkDeep learningArtificial intelligenceActivation function

Identifiers

DOI
10.3934/mfc.2023039
Journal
Mathematical Foundations of Computing
Year
2024