Course Description
This course provides a comprehensive introduction to deep neural networks. Major topics include multilayer perceptrons, convolutional neural networks, sequence modeling with recurrent neural networks and transformers, practical aspects of training deep neural networks, and generative probabilistic modeling with deep neural networks. Students will learn basic concepts of deep learning and hands-on experience to solve real-life problems. This course requires a strong background in linear algebra, probability and statistics, and machine learning. Python will be used for all the assignments.
Optional Textbooks
Deep Learning: Foundations and Concepts
Probabilistic machine learning
Deep Learning Frameworks
Communication
All written communication should be directed through Piazza. Sign-up instructions will be sent to your email. You can post publicly or privately depending on your preference. Emails won’t be responded to. Except for the sign-up phase of the class, we will not be using D2L for anything else.
Assignments
We will be using GitHub Classroom for all the assignments in this course. Sign-up instructions will be sent to your email.
Course Policy
Details on course policies can be found here.
Navigating the Lecture Slides
- The slides with audio work best on Safari, Chrome, Firefox, and Opera. I have not tested other browsers.
- If you want a PDF of the slides, press the ‘e’ key. It should create a PDF that can be downloaded.
- Hit the play button at the bottom of the screen. The slides should play the audio and auto advance.
- On Chrome, you can select playback speed in the audio control bar at the bottom.
- Press ‘n’ to manually navigate forward and ‘p’ to navigate backward. Do not use the left/right arrow keys.
- You can skip to any slide you like and press the play button at the bottom to listen to the audio for that slide.
- If the audio gets stuck or does not play, refresh that page. This is a known problem. It occurs when the browser does not fully load the audio file.
Tentative Schedule and Syllabus
Date | Lecture | Misc |
---|---|---|
Introduction | ||
Tue Jan 09 | No Class | |
Thu Jan 11 | Welcome (recorded) | |
Deep Networks | ||
Tue Jan 16 | Feed Forward Networks: Introduction (recorded) | |
Thu Jan 18 | Feed Forward Networks: Learning (recorded) | |
Tue Jan 23 | Backpropagation | |
Thu Jan 25 | Reverse Mode Automatic Differentiation | |
Tue Jan 30 | Optimization | |
Supervised Learning | ||
Thu Feb 01 | Convolutional Neural Networks | |
Tue Feb 06 | CNN Architectures | |
Thu Feb 08 | Interpretability (recorded) | |
Tue Feb 13 | No Class (university holiday) | |
Thu Feb 15 | Modeling Sequences | |
Tue Feb 20 | Attention | |
Thu Feb 22 | Transformers | |
Tue Feb 27 | No Class (Spring Break) | |
Thu Feb 29 | No Class (Spring Break) | |
Tue Mar 05 | State Space Models | |
Thu Mar 07 | How to Tame Your Deep Neural Network | |
Tue Mar 12 | Graph Neural Networks - I | |
Thu Mar 14 | Graph Neural Networks - II | |
Generative Models | ||
Tue Mar 19 | Generative Models: Introduction | |
Thu Mar 21 | Variational Autoencoders | |
Tue Mar 26 | Diffusion Models-I | |
Thu Mar 28 | Diffusion Models-II | |
Tue Apr 02 | Diffusion Models-III | |
Thu Apr 04 | Diffusion Models-IV | |
Assorted Topics | ||
Tue Apr 09 | Fairness in AI | |
Thu Apr 11 | Course Review | |
Tue Apr 16 | No Class | |
Thu Apr 18 | No Class |