Coding for cognition

Cognitive scientist Sacha Barber has created a three-part guide to the mechanics and mathematics of neural networks.

If you’re interested in how many artificial intelligence systems work, the guide takes you through the mathematics of neural networks, to the basic of creating your own network in the C programming language.

Even if you’re not a mathematician or programmer, the article is worth a scan so you can get an idea of the level of complexity that is needed for a group of mathematical functions to start displaying ‘cognitive’ properties.

The first thing you’ll notice is how simple the functions are.

With many sorts of neural networks, the difficulty is not in creating the ‘building blocks’ – i.e. the simulated neurons, but in creating a network structure that is useful for solving problems.

While engineers might be interested in creating networks to solve practical problems, neuropsychologists often create networks to simulate encapsulated mental processes, and then damage the networks to simulate brain damage.

This allows the researchers to test out ideas about how cognitive processes might be organised in the brain.

Link to ‘AI : Neural Network for beginners’.