I am new to BP, I am started learning it recently, unable to get the use of Bias node in BP...
Please explain it simply...
Jump to content
GameDev.net Posting Guidelines (please read before posting)
For Beginners Forum FAQs (please read before posting)
Subscribe to GameDev.net Direct to receive the latest updates and exclusive content.
Guest, the last post of this topic is over 60 days old and at this point you may not reply in this topic. If you wish to continue this conversation start a new topic.
Posted 11 March 2013 - 10:43 AM
By BP, what do you mean please?
Are you talking about back-propagation ? Yes you are, I didn't see your tag
If so, the bias is not related to back-propagation, but to Mc Culloch & Pitts model.
Any network model that use Mc Culloch & Pitts model must use a bias, as it is described in any good explanation.
The bias permit to shift the activation (or transfert) function, along x axis.
The common way to implement it is to add a constant input (1 or -1 usually) in input vector of each node and make its weight (i.e the bias) evolve like any other weight.
An analogy in 2D : the bias is like c in the line equation a.x+b.y+c=0
Edited by Tournicoti, 11 March 2013 - 03:04 PM.
GameDev.net™, the GameDev.net logo, and GDNet™ are trademarks of GameDev.net, LLC.