Procedural learning

From supermemo.guru
Jump to navigation Jump to search

Procedural learning is learning in which a neural network optimizes its connections with the help of reward and penalty signals. In opposition to declarative learning, the network does not get ready answers or correct associations on input. All it gets is valuations of its outputs that change upon minor random or conscious adjustments. Procedural learning is best known from motor tasks such as riding a bike or playing a piano.

While riding a bike, we try to influence the motor system networks consciously at first. However, gradually, we switch to continual improvement by just riding. Minor random fluctuations can occur in the network at any time, however, the performance stabilizes asymptotically, and the conscious control is gradually withdrawn. Random fluctuations also decrease due to an increase of synaptic stability in the system.

This glossary entry is used to explain "I would never send my kids to school" (2017-2024) by Piotr Wozniak