This repository contains two major work paths: one to evaluate, under a beginner's point of view, four frameworks used for Deep Learning through an implementation of a neural network to recognise handwritten digits from the MNIST database (you can see our work on MNIST here) and another to use our evaluation to choose a framework and utilize a language representation (namely word embeddings) to classify medical papers into different categories (see Epistemonikos folder).