This course offers an introduction to data science as well as various software tools. It provides a comprehensive presentation of neural networks: deep, convolutional, recurrent, adversarial and generative. It also provides an introduction to the tools routinely used by data analysis practitioners. An important part of the course is devoted to practical case studies on computers, using Jupiter notebooks. More specifically, we will study the categorization of images, semantic segmentation of images and speech recognition. Part of the evaluation will take the form of a Kaggle challenge.
Syllabus:
Introduction: methodological framework, introduction to Kaggle and Python Notebooks.
Convolutional Neural Networks (CNNs).
Recurrent Neural Networks.
Adversarial and generative networks.
IBM Data Science tools.
Planning & teachers:
18 septembre 13h30 2h CM 2h Lab Lionel Fillatre
25 septembre 13h30 2h CM 2h Lab Pierre Alliez + Gaetan Bahl
02 octobre 13h30 2h CM 2h Lab Florent Lafarge + Gaetan Bahl
09 octobre 13h30 2h CM 2h Lab Pierre Alliez + Gaetan Bahl
16 octobre 16-19h Yann Gouedo
23 octobre 16-19h Yann Gouedo
30 octobre 17-19h Yann Gouedo
06 novembre exam Pierre Alliez