ML Anthology

Analysis of Neural Networks with Redundancy

Yoshio Izui, Alex Pentland
NeCo 1990 pp. 226-238
doi:10.1162/NECO.1990.2.2.226 /neco/1990/izui1990neco-analysis/

Abstract

Biological systems have a large degree of redundancy, a fact that is usually thought to have little effect beyond providing reliable function despite the death of individual neurons. We have discovered, however, that redundancy can qualitatively change the computations carried out by a network. We prove that for both feedforward and feedback networks the simple duplication of nodes and connections results in more accurate, faster, and more stable computation.

NeCo Semantic Scholar

Cite

Text

Izui and Pentland. "Analysis of Neural Networks with Redundancy." Neural Computation, 1990. doi:10.1162/NECO.1990.2.2.226

Markdown

[Izui and Pentland. "Analysis of Neural Networks with Redundancy." Neural Computation, 1990.](https://mlanthology.org/neco/1990/izui1990neco-analysis/) doi:10.1162/NECO.1990.2.2.226

BibTeX

@article{izui1990neco-analysis,
  title     = {{Analysis of Neural Networks with Redundancy}},
  author    = {Izui, Yoshio and Pentland, Alex},
  journal   = {Neural Computation},
  year      = {1990},
  pages     = {226-238},
  doi       = {10.1162/NECO.1990.2.2.226},
  volume    = {2},
  url       = {https://mlanthology.org/neco/1990/izui1990neco-analysis/}
}