Deep Learning

Course Description

In many real-world Machine Learning tasks, especially those with perceptual input, such as vision and speech, the mapping from raw data to the output is often an overly complicated function with many factors of variation. In the past, to achieve acceptable performance on such tasks, significant effort had to be expended to engineer hand crafted features. However, with the advent of deep learning, such tasks have been made easily facilitated and more realistic. In this respect, deep Learning algorithms aim to learn feature hierarchies with features at higher levels in the hierarchy formed by the composition of lower-level features. This automatic feature learning has been demonstrated to uncover underlying structure in the data leading to state-of-the-art results in tasks in vision, speech and rapidly in other domains as well. This course centers on learning the basic theory of deep learning and how to apply it to various applications. It aims to present the mathematical, statistical, and computational challenges of building stable representations for high-dimensional data, such as images, text, and data. It will delve into selected topics of Deep Learning, discussing recent models from both supervised and unsupervised learning. Special emphasis will be on convolutional architectures, invariance learning, unsupervised learning, and non-convex optimization and other related topics in the area of deep learning.

Course ID: ARTI 406