Training feedforward neural network via multiobjective optimization model using non-smooth L1/2 regularization | Publicación