Hand-coded neural networks
Here’s a JavaScript function that
returns a positive value if both x and y are positive:
function bothPos(x, y) {
const negX = Math.max(0, -x);
const negY = Math.max(0, -y);
return Math.max(0, 1-negX-negY);
}
Surprise - this function is also a neural network! Neural networks are usually drawn as
No, the neural network is not a series of neurons; it’s a JavaScript function which uses a restricted set of operations.
It approximates the function below, which is perhaps how you might naturally implement it:
function bothPos(x, y) {
return x > 0 && y > 0 ? 1 : 0;
}
Hear “neural network” and you probably think “machine learning”. But it’s useful to think of neural networks independently of machine learning.
Tagged #programming, #machine-learning. All content copyright James Fisher 2019. This post is not associated with my employer.
TigYog