Deep Learning Demo🔗
Evaluation 2025-2026
Principle and Pedagogical Objective🔗
The purpose of this exercise is to create an educational video and code demonstration presenting the topic, as if you were presenting it to your peers (SDD students, work colleagues, knowledgeable clients, etc.). We never learn as well as when we teach, so this is an opportunity to master one more topic and collectively build competence by critically reviewing each other's work. We will use peer evaluation, and the goal is for the time you dedicate to this assessment to be a time where you continue to discover new things and deepen your mastery of deep learning. You will choose a specific topic to present, either from the list of topics or of your choice, if your proposed subject is validated.
Starting Point🔗
Each topic is associated with one or two online references. These are starting points, not content to summarize! It is expected that you will read complementary documents, build your understanding, and make choices about what to present in your video and demo (or in some cases, choose to specifically address one aspect of the topic and not another - in this case, please discuss it with me).
Instructions and Evaluation Criteria🔗
Video Requirements🔗
Your video must be educational, engaging, and have a good balance between formal and practical aspects. The suggested length is 10 minutes, with a minimum of 7 minutes and a maximum of 12 minutes. This is sufficient time to fully cover deep learning concepts if the video is well planned.
You can use OBS, Camtasia, DaVinci Resolve, or other tools to create your video. You are encouraged to plan out your video and create slides to make it truly educational. The language can be French or English, according to your preference (language quality contributes to a pleasant viewing experience and to your grade).
Your video must be rigorous: this is not scientific popularization, you must be precise and rigorous. This doesn't require you to systematically provide mathematical proofs, but it does require formulating and discussing ideas and results in a precise and well-argued manner.
Your video must clearly explain the deep learning concepts in the article, making them accessible and understandable to your peers.
Demo Requirements🔗
Your code demonstration must be useful and reusable, featuring elements that will allow viewers to quickly become functional with the topic. The demo should provide results and support for the concepts explained in your video.
Ease of use is important: other students will try to run your demo, so make sure it is well-documented with clear installation and execution instructions (e.g., a README file with dependencies, setup steps, and usage examples).
Your demo must be well-documented: code should be clear, and any necessary references or resources should be included.
Extension and Originality🔗
Beyond reproducing the results or concepts from the article, you must explore original ideas. How can you improve on the idea or modify it? Even if your modification doesn't improve performance, exploring alternative approaches, testing hypotheses, or applying the concept to new domains demonstrates deep understanding and creativity.
This exploratory component is an essential part of the assignment and will be evaluated based on the originality and execution of your extension idea.
Evaluation Criteria🔗
| Criterion | Points |
|---|---|
| Video clarity and pedagogy | 5 |
| Video comprehensiveness - technical accuracy, quality, and level of detail | 5 |
| Demo code quality and ease of use | 5 |
| Extension idea originality and execution | 5 |
| Total | 20 |
Submission🔗
You must submit: 1. Video: Upload to YouTube, Vimeo or Nextcloud 2. Code Demo: Upload to GitHub
Submission process: Via the LMS (Learning Management System), submit only the links to both your video and your GitHub repository. The LMS link will be provided separately.
Schedule🔗
| Date | Event |
|---|---|
| 05/01/26 | Subject selection |
| 30/01/26 | Submission deadline |
| 02/02/26 | Peer evaluation |
It is important to submit the links beforehand to ensure that everyone has a video and demo submitted before the peer evaluation.
Peer Evaluation🔗
After submission, other students will evaluate your work by: 1. Watching your video 2. Running your demo
They will assess the clarity of your explanations, the technical depth, the quality and usability of your demo, and the originality of your extension work. This peer evaluation process is an opportunity to learn from each other's approaches and to provide constructive feedback.
Work Recommendations🔗
Creating a quality video and demo takes time, but it's also one of the best ways to learn deeply. Here's a recommended workflow:
Week 1: Research sources and efficient reading. Each topic has bibliographic references - read them efficiently and search for complementary sources to better understand or gain different perspectives.
Week 2: Deep reading and practical experience with the topic (coding, personal exploration of theory).
Week 3: Decide on your video structure, create slides, and draft your script. Begin developing your demo code.
Week 4: Record your video and finalize your demo. Review, polish, and test everything. Make sure your demo runs smoothly and is well-documented.
Your CV🔗
This video and demo are part of your portfolio. Put them on your GitHub, mention them in interviews, and highlight them when you apply somewhere. This is a unique and valuable exercise that is completely worth showcasing in your CV! Both the video and demonstration code can be excellent additions to your professional portfolio.