Neural networks are a useful alternative to Gaussian mixture models for acoustic modeling; however, training multilayer networks involves a difficult, nonconvex optimization that...
The task of finding the optimum of some function f(x) is commonly accomplished by generating and testing sample solutions iteratively, choosing each new sample x heuristically on t...
— Nowadays, huge amounts of information from different industrial processes are stored into databases and companies can improve their production efficiency by mining some new kn...
"Gaussian processes (GPs) provide a principled, practical, probabilistic approach to learning in kernel machines. GPs have received increased attention in the machine-learning...
Carl Edward Rasmussen and Christopher K. I. Willia...
In the Neural Networks approach by Radial Basis Function - RBF, the property of interpolation between faces, their variation, and the diversity of faces helps to minimize the outp...
Antonio C. Zimmermann, L. S. Encinas, L. O. Marin,...