DLAI-s2-2023
Deep Learning & Applied AI @Sapienza
Course material, 2nd semester a.y. 2022/2023, Dept. of Computer Science
News
- 01/08/2023: The grades for the written exam of July 19th are available here
- 29/06/2023: The grades for the written exam of June 27th are available here
- 11/05/2023: The list of projects is now published, please scroll down to the Grading section for more details.
- 10/05/2023: Due to logistic issues, today’s lecture will be held remotely through a pre-recorded video; please scroll down to download the video. Apologies for the inconvenience, and please feel free to reach out to the Professor if you have any questions.
- 07/05/2023: Please fill out the OPIS questionnaire (instructions here). The code for this course is 2YSAJJWX.
- 07/05/2023: The midterm grades are now published.
- 26/04/2023: The midterm sheet is now published. Students who wish to have their answers checked must deliver by today at 15:15 via email (rodola@di.uniroma1.it).
- 19/04/2023: The lecture of April 25th is cancelled due to the Liberation Day, as per the academic calendar. On April 26th we will do a self-evaluation test; more details will follow.
- 03/04/2023: The lecture of April 11th is cancelled due to Easter holidays, as per the academic calendar.
- 27/03/2023: The lectures of Tue 28 and Wed 29 March will be held remotely due to illness. Tue 28: The notebook will be uploaded shortly, and you can work on it from your homes; the Professor and the teaching assistants will be on the Discord server during the 3 hours, to assist you in real time. Wed 29: The slides are already up, together with a video recording of the Professor explaining the new topic; the recording is from DLAI 2021, which overlaps well with the new slides up to some simplifications.
- 11/02/2023: The course website is online. Welcome to DLAI 2022/23! The course will start on Tue 28th February.
Logistics
Lecturer: Prof. Emanuele Rodolà
Assistants: Dr. Antonio Norelli, Dr. Luca Moschella, Dr. Marco Fumero, Dr. Irene Cannistraci
When: Tuesdays 13:00–16:00 and Wednesdays 13:00–15:00
Where:
Physical classrooms: Aula Magna Edificio C RM111 (Tuesdays) and Aula Alfa RM062 (Wednesdays)
There is no virtual classroom, and the lectures will not be recorded.
Q & A: We will use a Discord server. More details during the first lessons.
Pre-requisites
Python fundamentals; calculus; linear algebra.
Textbook and reading material
Due to the continuously evolving nature of the topic, there is no fixed textbook as a reference. Specific material in the form of scientific articles and book chapters will be given throughout the lectures.
In addition, here you can find some supplementary course notes.
Grading
Evaluation proceeds according to the following steps:
- A midterm self-evaluation test (optional, does not concur to the final grade)
- A final written exam (mandatory, accounts for 60% of the final grade)
- A project (mandatory, accounts for 40% of the final grade)
- An oral exam (optional, attributes at most 3 points, added to or subtracted from the final grade)
We may require an oral exam in doubtful cases or whenever necessary.
The list of projects is available here. Each project must be accompanied with code + a 2 page report using a fixed template, available here. Projects can be made in groups of at most 2 students, but in this case, you must motivate this decision and get our approval beforehand.
Here you can find some example sheets of past written exams:
Lectures
| Date | Topic | Reading | Code & Data |
|---|---|---|---|
| Tue 28 Feb | Introduction | slides | |
| Wed 01 Mar | Data, features, and embeddings | slides ; linear algebra recap ; matrix notes | |
| Tue 07 Mar | Tensor manipulation and Tensor operations |  |
|
| Wed 08 Mar | Linear regression, convexity, and gradients | slides | |
| Tue 14 Mar | Linear models and Pytorch Datasets | |
|
| Wed 15 Mar | Overfitting and going nonlinear | slides | |
| Tue 21 Mar | Logistic Regression and Optimization | |
|
| Wed 22 Mar | Stochastic gradient descent | slides | |
| Tue 28 Mar | Autograd and Modules | |
|
| Wed 29 Mar | Multi-layer perceptron and back-propagation | slides ; video | |
| Tue 04 Apr | Convolutional Neural Networks | |
|
| Wed 05 Apr | Convolutional Neural Networks | slides | |
| Tue 11 Apr | Easter holidays | ||
| Wed 12 Apr | Regularization, batchnorm and dropout | slides | |
| Tue 18 Apr | Uncertainty, regularization and the deep learning toolset | slides | |
| Wed 19 Apr | Deep generative models | slides ; video | |
| Tue 25 Apr | Liberation Day | ||
| Wed 26 Apr | Midterm self-evaluation | sheet ; grades | |
| Tue 02 May | Variational Autoencoders | |
|
| Wed 03 May | Geometric deep learning | slides; video | |
| Tue 09 May | Self-attention and transformers | slides | |
| Wed 10 May | Adversarial training | slides ; video | |
| Tue 16 May | CycleGAN and Adversarial Attacks | |
|
| Wed 17 May | Invited lecture: Andrea Santilli - “From symbolic representations to ChatGPT” | slides | |
| Tue 23 May | Invited lecture: Michele Mancusi and Giorgio Mariani - “Diffusion-based generative models for audio” | slides | |
End