There are multiple very complicated articles about softmax online. I just want to know a few things about it:

  1. Why do we need softmax? What is the "problem"? What does it do?
  2. How does it do it?

I find it very hard to get through all the math when I don't really get what it is all about. The problem is that all the things I read online is very complicated written and no one seems to just get to the point.

I hope you can help.

  • 1
    $\begingroup$ There are also many simple articles. What research did you do? What is the simplest article you read, and what did you understand and what did you not understand? It looks like your questions are answered on Wikipedia: en.wikipedia.org/wiki/Softmax_function. I don't see much point in repeating information that is already widely available in standard resources. It may be hard to fully explain a mathematical function without using math. $\endgroup$
    – D.W.
    Nov 23 '20 at 23:56

Ok, I have read a very good explanation to what softmax is.

A dog is programmed to walk through a maze. One start point has very few walls. The way it is programmed with the epsilon greedy algorithm, the dog will sooner or later start to walk mostly to the right or mostly to the left. That is a problem because it at some point might just walk in a circle the entire time. To get that out one could make it choose left or right randomly from time to time - but that is not really a good approach for obvious reasons.

So instead you use something called softmax. Softmax just smoothens the decision process: like taking out the walking-left or walking-right.


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.