ML Anthology

Precision and Approximate Flatness in Artificial Neural Networks

Maxwell B. Stinchcombe
NeCo 1995 pp. 1021-1039
doi:10.1162/NECO.1995.7.5.1021 /neco/1995/stinchcombe1995neco-precision/

Abstract

Several of the major classes of artificial neural networks' output functions are linear combinations of elements of approximately flat sets. This gives a tool for understanding the precision problem as well as providing a rationale for mixing types of networks. Approximate flatness also helps explain the power of artificial neural network techniques relative to series regressions—series regressions take linear combinations of flat sets, while neural networks take linear combinations of the much larger class of approximately flat sets.

NeCo Semantic Scholar

Cite

Text

Stinchcombe. "Precision and Approximate Flatness in Artificial Neural Networks." Neural Computation, 1995. doi:10.1162/NECO.1995.7.5.1021

Markdown

[Stinchcombe. "Precision and Approximate Flatness in Artificial Neural Networks." Neural Computation, 1995.](https://mlanthology.org/neco/1995/stinchcombe1995neco-precision/) doi:10.1162/NECO.1995.7.5.1021

BibTeX

@article{stinchcombe1995neco-precision,
  title     = {{Precision and Approximate Flatness in Artificial Neural Networks}},
  author    = {Stinchcombe, Maxwell B.},
  journal   = {Neural Computation},
  year      = {1995},
  pages     = {1021-1039},
  doi       = {10.1162/NECO.1995.7.5.1021},
  volume    = {7},
  url       = {https://mlanthology.org/neco/1995/stinchcombe1995neco-precision/}
}