• Advertisement
Sign in to follow this  

Making the activation of a NN in between x&y rather than above x?

This topic is 4243 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

If you intended to correct an error in the post then please contact us.

Recommended Posts

I don't know anything about the mathametical theory behind NNs so this might be a retarded idea for those that do. But suppose the criteria for the activation of a neuron wasn't that a number be greater than X but that it fall between X and Y? Would the network learn faster or just not learn at all?

Share this post


Link to post
Share on other sites
Advertisement
Why no replies? I guess I can just run a NN myself and find out but I was hoping someone on here could save me the trouble.

Share this post


Link to post
Share on other sites
isn't that the same as combineing a nn node with a trigger when below x and a node with a trigger on high y into the output?

Share this post


Link to post
Share on other sites
Well - running the net yourself is never a bad idea - if only for your own interest.

However what you are asking is basically covered here: http://www.gamedev.net/community/forums/topic.asp?topic_id=116850

I am to lazy to retype it all. Basically if you are using a gradiant based learning method you need to use a Heaviside function for the activation.

Hope this helps :)

Share this post


Link to post
Share on other sites
Sign in to follow this  

  • Advertisement