Perceptron
device · 1958 · Digital age
The first trainable neural network machine (Rosenblatt).
Inventor: Frank Rosenblatt.
What it is made of
- Potentiometer — has part. Potentiometers set the weights.
- Photodiode — has part. Photocells read the image.
- Electric motor — has part. Motors adjusted the weights during training.
Principles and materials it uses
- Copper — is made of. The Mark I is wires and potentiometers.
- Artificial neural network — uses. The first trainable neural network.
What it is a kind of
- Artificial neural network — is a kind of. The first neural network.