Abstract: “In this talk Dr. Héctor Allende-Cid, professor of the Pontificia Universidad Católica de Valparaíso, Chile, will present 3 cases of his current work of Machine Learning applied to Natural Language Processing tasks. The first one, is based in the classification of schizophrenic patients using transcribed narrative texts. In this work, two main objectives are pursued: using interpretable classification algorithms to describe the patterns that differentiate schizophrenic and control patients, and to achieve the best classification performance measures using Machine Learning algorithms. In the second work, Dr. Allende-Cid will present ongoing work in the task of Authorship Attribution, by performing classification tasks over feature vectors of POS-tagging networks generated from the analyzed texts. The last case deals with Sentiment Analysis of Political Tweets in the context of the 2017 Chilean Presidential election, comparing classical ML algorithms and representations with Deep Recurrent Neural networks and Word2vec.”