Talk:Artificial Neural Networks/Activation Functions

In the section titled "Continuous Log-Sigmoid Function" the definition of sigma(t) is inconsistent with the derivative given. I think that either the definition of sigma(t) should have the parameter beta removed (set to 1), or (preferably) the derivative should include a factor of beta. Currently the derivative given for sigma(t) is not accurate. 75.163.116.16 (talk) 17:40, 26 September 2010 (UTC)