What do machines think about?

Abstract

Can machines think? This question was posed by Alan M. Turing (1950) in the mid-20th century. The answer to that question is the proposal of the so-called Turing test. In this test, Artificial Intelligence (AI) is considered to be a way of acting that imitates the intelligent behavior of human beings. Since then the AI has been surpassing the human being in tasks for which it was supposed to have intelligence: strategy games like chess, driving vehicles, composing symphonies, automatic planning, and a long etcetera that seems to have no end. In fact, the changes produced in the last decades in the telecommunications sector, accompanied by the development of the storage and processing capacity of information have meant a change of paradigm to which the name Industry 4.0 has been given. AI corresponds to a field of knowledge that includes Machine Learning (ML) and Deep Learning (DL). In both fields, to solve a problem proceeds to the training of models to learn the problem in question from existing data. Once the rules are obtained, we can apply them to new data sets to produce the appropriate answers by applying the rules learned from experience. To perform ML processes at least three fundamental parts are necessary: input data, the expected results and the measurement of the algorithm's performance so that the algorithm's work can be adjusted by means of feedback processes.