Towards the Formalization of a Trustworthy AI for Mining Interpretable Models explOiting Sophisticated Algorithms