“…[8,9]. However, even if some applications of neural networks to predict metals flow stress have been proposed in the last years [6,7,10,11], very few studies are recordable in scientific literature that utilise neural networks to describe material behaviour under varying hot deformation conditions. Material response in terms of flow strength or microstructure is predicted in case of well-defined deformation conditions, mainly single-step deformation conditions; but, even if these operative conditions are near to those of conventional rheological testing, they can be very different from those of industrial hot forging operations [1,2].…”