PhD Course | From 17 June 2024 to 19 June 2024

Stochastic Deep Learning Models in Space and Time for Environmental Applications


Prof. Stefano Castruccio
Department of Applied and Computational Mathematics and Statistics
Notre Dame University, IN, USA

Course Description
This course is intended to introduce graduate student to mathematical and computational foundations of deep neural networks. The students will learn what has made deep neural networks extremely successful over the past two decades and how to properly implement them in Python. There will be both lectures and in class code tutorial, and the main focus of the course will be on implementing neural networks on practical problems, with research applications in both engineering and science.
The course introduces the mathematical and computational background to use deep neural networks. The students will be introduced to Python and Google Colab, and neural networks will be presented as generalization of linear regression models. Fundamentals of parameter learning and image recognition will also be covered.

Course structure

  • Introduction to Python and Google Colab
  • Review on linear regression
  • Introduction to neural networks (NNs)
  • Learning networks: gradient descent and Backpropagation
  • Image recognition and convolutional NNs

Notebook with Python and Google Colab installed and tested.

Locandina Stochastic Deep Learning Models in Space and Time.pdf
200 KB