2
$\begingroup$

I am working on some content to teach ANN's to gradeschool kids. I've run across a couple good articles, most of which model a simple XOR problem, using three layers (input, 1 hidden, output).

however, this article is a little different in that it simply goes straight from inputs to outputs, with no hidden layer in between. Most other articles seem to follow the three layer approach for hello world ANN models.

Does this technically qualify as an ANN for teaching purposes, or must we have at least three layers to be an ANN? (input, 1 hidden, output). I want to use this simpler model as the first ANN in the content I am producing, but...but only if it technically qualifies as a neural net.

Thanks!

$\endgroup$
3
$\begingroup$

It is a definition issue. Different people have different definitions on what is ANN.

For one single layer network, people always use the term perceptron in ANN terms. And it has very close relationship to logistic regression.

See this question for details

What's the difference between logistic regression and perceptron?

The term perception indicates there are connections between the model with biology.

Here is a tutorial may fit your use.

https://appliedgo.net/perceptron/

And it defines perception as ANN with biological connections.

$\endgroup$
  • $\begingroup$ thank you @hxd1011. This answer is perfect, very helpful, especially the tutorial link. $\endgroup$ – 3z33etm Oct 16 '17 at 17:05

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.