Deep LearningPermalink

ObjectivesPermalink

  1. Define the main concepts in the fields of Machine Learning, focusing on the area of Deep Learning.
  2. Know appropriate classes of methods / algorithms, applications and programming libraries for solving the main problems in Deep Learning.
  3. Apply available Deep Learning software for problem solving, including the use of open source libraries
  4. Build programs that can use available software libraries for the implementation of advanced deep learning pipelines including supervised, unsupervised and reinforcement paradigms.
  5. Build programs that can implement existing algorithms or develop new deep learning algorithms.

ProgramPermalink

  1. Linear and nonlinear functional models of supervised learning: linear regression and logistics, Support Vector Machines
  2. Neuronal Networks: functioning of a neuron, feedforward networks, training algorithms
  3. Deep Learning: Architectures; training algorithms; regularization and dropout; multi-task learning
  4. Convolutional Neuronal Networks: Convolutions; layer types and architectures; image / video processing applications; transfer learning; convolutional graph networks and their applications
  5. Unsupervised and semi-supervised learning: embeddings, tSNE data visualization, deep learning-based clustering, auto-encoders
  6. Recurrent Neuronal Networks: layer types and network architectures; networks with memory; attention mechanisms; applications
  7. Generative and adversarial deep learning models: variational auto-encoders, generative adversarial networks, applications
  8. Deep reinforcement learning approaches and their applications

BibliographyPermalink

I. Goodfellow, Y. Bengio, A. Courville. Deep learning. MIT Press, 2016. W. Richert, L.P. Coelho. Building machine learning systems with python. Packt publishing. 2013. T. Mitchell, Machine Learning, McGraw Hill, 1997. F. Chollet. Deep Learning with Python. 2017

Updated: