Group for Research in Decision Analysis

Introductory Data Science and Artificial Intelligence Workshop using Python

Vahid Partovi Nia Researcher, Department of Mathematical and Industrial Engineering, Polytechnique Montréal, Canada

Mouloud Belbahri Université de Montréal, Canada

David Berger Université de Montréal, Canada

Gaetan Marceau Caron MILA, Canada


You must register to participate in the workshop. Please note that the number of participants is limited and lunch will not be provided.

If the maximum number of participants is reached, please contact Vahid Partovi Nia, indicating that you are interested in the workshop. Another workshop may be available later.

A data scientist requires a set of skills for extraction of knowledge from large volume of data. Deep learning is a sub-field of artificial intelligence, aiming at discovering the underlying pattern of data through predictive models.

This one-day workshop is interactive, and you require to bring your laptop with anaconda, python, and jupyter-notebook installed on, in order to code with us during the workshop. There is no python familiarity required for this workshop. Being equipped with some basic computer programming skills helps. There is no statistics or mathematics knowledge required, however some knowledge in the introductory statistics level helps.

The aim of this workshop is to introduce some elementary data science and artificial intelligence concepts such as data loading, data visualization, supervised and unsupervised learning, neural networks, and deep learning.

We start with some basic Python libraries such as numpy, scipy, matplotlib, and move towards machine learning libraries such as scikit-learn and keras. You will learn how to execute several supervised, unsupervised, and semi-supervised learning algorithms on jupyter notebook using python 2.7.

We will go over the details of some important tools that are widely used in data science and artificial intelligence, along with the algorithm background, some preliminary mathematical insights, and some statistical intuition.

As a data scientist you must be able to communicate or collaborate with potential data-related projects effectively, using powerful programming tools. Python is not only famous because of its computational power, but also because of its easy integrity in web and mobile applications.

Python codes are easily readable and effectively maintainable. It is an easy-to-understand language developed by computer engineers which can be learned by experts from various backgrounds such as statistics, mathematics, business, finance, etc.

This workshop enables you to build a comprehensive tool that unifies every part of your data-driven workflow. Python and its libraries are platform independent, and are run on Windows, Linux, or Mac OS X identically. MILA research lab develops its deep learning library "theano" on python. However, for this course we use a simpler interface named "keras". "keras" is built over "theano".

This workshop is partly sponsored by IVADO and GERAD.