Lecture 1

In the first lecture, we will introduce the course and its goals. We will describe the projects, the evaluation metrics, the prerequisites.

Lecture 2

In this lecture, we will introduce the GNU/Linux operating system and we will describe its main command. We will then describe how to use a HPC cluster.

Lecture 3

In this lecture, we will introduce Git which is a fundamental version control system often use for coordinating work among programmers.

Lecture 4

In this lecture, we will summarize the most popular deep learning models used for processing speech signals. We will start with a recap of of basic deep learning algorithms and we will then refresh convolutional and recurrent neural networks. We will conclude with a description of transformers.

Lecture 5

In this lecture, we will introduce the basics of speech processing. We will talk about speech analysis, speech representations, phonemes and their features. The goal of this lecture is to prepare the students for their speech processing projects.

Lecture 6

In this lecture, we will discuss some tips and trick useful when working on a deep learning project (e.g, data augmentation, management of OOM, numerical instabilities, hyperparameter tuning, overfitting vs underfitting, sanity checks, learning rate scheduling…).

Lecture 7

In this lecture, we will provide a tutorial on advanced PyTorch functionalities. Note that a basic PyTorch tutorial will be given before in the "IFT 6135 - Representation Learning" course. We encourage the students of both courses to follow the basic and advanced pytorch tutorials.

Lecture 8-9

In this there lectures, we will describe basic and advanced functionalities of SpeechBrain. We will show how to implement speech recognition, speech enhancement, and separation systems. We will also provide a tutorial on multi-microphone signal processing.