1. A Model for Neural Network Modeling in Neuroscience.
- Author
-
Storrs, Katherine R. and Maiello, Guido
- Subjects
- *
ARTIFICIAL neural networks , *SIGNAL convolution , *CONVOLUTIONAL neural networks , *NEUROSCIENCES , *PATTERN recognition systems , *MOTION perception (Vision) - Abstract
The article focuses on MotionNet, a convolutional neural network comprised of layers of interconnected units that have spatially restricted receptive fields and compute weighted sums of their inputs. The fully transparent, fully interrogable artificial model was developed and described by R. Rideaux and A. E. Welchman in their research published in the 2020 issue of this publication. The biological computations underlying motion perception are explored using MotionNet.
- Published
- 2020
- Full Text
- View/download PDF