- Teaching Staff:
- Projects
- 🔵 List of classes - including lecture slides and videos
- Notes and links
- Administration
- Supporting Material
Teaching Staff:
Lectures: Gal Chechik, gondaneuralnetworks @ gmail.com. TA: Shauli Ravfogel, shauli321 @ gmail.com
When and where:
- Zoom (Lectures, Tirgul. Password for the tirgul: nn-course).
- Mon. 10:00-11:30, Tue. 10:00-11:30 (Gal)
- Tue. 14:00-15:30 (Shauli)
Projects
Course projects can be done individually or in pairs (encouraged). Students should propose a project that involves applying one of the learning techniques covered in the course to non-trivial data. It is allowed and encouraged to use data from your MSc or PhD research.
Grading criteria for the project: (1) Proper formulation of the learning problem. (2) Adequacy of the method to the learning problem (3) Level of complexity and scale of the learning problem (4) Novelty of technical ideas. (5) Quality of presentation.
🔵 List of classes - including lecture slides and videos
🔵 Syllabus
Submission deadline | Files | Grades |
---|---|---|
2.11.20 | EX 1, part 2 | grades |
9.11.20 | EX 2 | grades |
16.11.20 | EX 3 | grades |
25.11.20 | EX 4 | - |
20.12.20 | EX 5 | - |
Notes and links
Slides from Tirgul (Shauli, 2020):
Note: the content of the slides may change significantly until the tirgul.
- Class 1: Introduction, projections, GD, Newton method Video
- Class 2: MLE, Logistic regression, colab Video
- Python intro video, Colab Example 1, Colab example 2
- Class 3: Multiclass, MLPs Video
- Class 4 - Backpropagation Video
- Class 5: SVMs, Deep Dream Video
- Class 6: Multivariate Gaussians, PCA Video
- Class 7: Yoav Goldberg’s Attention slides Video
- Class 8: VAE Video
- Class 9: VAE (Cont.) Video
- [Class 10] Torch Video
- [Class 11] Projects discussion Video
Administration
The final grade is computed as a combination of (1) home assignments (2) Final project. Students must complete all three to pass the course.
Grade = 0.5 * Home Assignments + 0.5 Final project.
Home assignment policies: Students are allowed to solve assignments in pairs. They must write the solutions themselves and submit each assignment as individuals. If you have collaborated when working on an assignment, mark clearly the name of the people you collaborated with.
How to submit:
Submit your colab notebook/a pdf with your written answers in the mail.
- submission: Email it to: nn.homework.2020 @ gmail.com by the deadline.
Acceptable format of written answers: PDF
The email title should include #assignment, and the submitting students full name and ID.
For example, Email subject: “Assignment 3, submitted by Alon Cohen, ID 01234787485”.
- Late home assignment submissions: Each late day results in a 5 points penalty, and the assignment will not be graded if submitted after 4 days or more.
Supporting Material
- ML: Machine Learning and Pattern recognition, C. Bishop (partially available online) (PDF)
- ML: The Elements of Statistical Learning, T. Hastie et al. (available online)
- ML: Machine learning course @ Stanford University online course (notes and videos)
- ML: Mathematical monk course on machine learning (video lectures)
- NN: Neural Networks course videos by Hugo Larochelle
- Understanding LSTMs
- NN: Deep Learning (Goodfellow’s book)