Dimensionality Reduction🔗

This class goes deeper in the mechanism and usage of Dimensionality Reduction. You will learn to exploit the effects of dimensionality on your machine learning models, with illustrated intuitions, and practical examples. You will start with an already very known technique in a linear setup, Principal Components Analysis, to move progressively towards non-linear manifold learning with t-SNE, to end up with Deep Learning techniques to reduce the dimension. Through concrete application of Autoencoders you will know what are their main interests and pitfalls, to make the best architecture choice given your problem to solve.