School LACSC: Deep neural networks: Fundamentals, computational, mathematical aspects, and advances

Deep neural networks (DNN) have boomed as a must-have technique in almost all machine-learning applications. After nearly a decade of evolution, many deep-based methods have been proposed considering more than only earlier convolutional operations in the hidden layers of an artificial neural network. The goal of this tutorial is not only to present the fundamentals of the computational and mathematical aspects of DNNs, considering from the very first proposed network of this kind - the Lenet-5 and its main components but also the more advanced architectures such as the Transformer networks that do not use convolution operations to do the job. All the technical aspects studied will rely on the digital image domain to exemplify and explain the fundamentals of the DNNs.

Register Now