What is the difference between training function and learning function

Question

What is the difference between Training function and learning function in Artificial Neural Network ? I have found this , but can't understand properly. Would anybody please explain ??

Answer 1

The training function is the overall algorithm that is used to train the neural network to recognize a certain input and map it to an output. A common example is backpropagation and its many variations and weight/bias training. A learning function deals with individual weights and thresholds and decides how those would be manipulated. These usually (but not always) employ some form of gradient descent. Examples include simulated annealing, Silva and Almeida's algorithm, using momentum and adaptive learning-rates, and weight-learning (examples include Hebb, Kohonen, etc.) algorithms.