Is there a way to specify some weights of ANN, as I build a parallel ANN structure, and the output of the whole ANN is the summary of sub-ANN, say Out=Out1+Out2-Out3,so the weight of the output neuron are 1,1,-1. My question is: can I specify these three weights, while the other weights will be updated automatically?

 Réponse acceptée

Greg Heath
Greg Heath le 3 Fév 2015

0 votes

Set the learning rates of the output layer to zero.

1 commentaire

liang
liang le 3 Fév 2015
That sounds make sense, but I do not find how to set the lr of the output layer to zero.

Connectez-vous pour commenter.

Plus de réponses (0)

Catégories

En savoir plus sur Deep Learning Toolbox dans Centre d'aide et File Exchange

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by