Study programme 2023-2024 | Français | ||
First-Order Methods for Large Scale Machine Learning | |||
Programme component of Master's in Computer Engineering and Management (MONS) (day schedule) à la Faculty of Engineering |
Code | Type | Head of UE | Department’s contact details | Teacher(s) |
---|---|---|---|---|
UI-M2-IRIGIG-204-M | Compulsory UE | GILLIS Nicolas | F151 - Mathématique et Recherche opérationnelle |
|
Language of instruction | Language of assessment | HT(*) | HTPE(*) | HTPS(*) | HR(*) | HD(*) | Credits | Weighting | Term |
---|---|---|---|---|---|---|---|---|---|
| Anglais | 8 | 16 | 0 | 0 | 0 | 2 | 2.00 | 2nd term |
AA Code | Teaching Activity (AA) | HT(*) | HTPE(*) | HTPS(*) | HR(*) | HD(*) | Term | Weighting |
---|---|---|---|---|---|---|---|---|
I-MARO-303 | First-Order Methods for Large Scale Machine Learning | 6 | 12 | 0 | 0 | 0 | Q2 | 75.00% |
I-MARO-304 | First-Order Methods for Large Scale Machine Learning - Complements | 2 | 4 | 0 | 0 | 0 | Q2 | 25.00% |
Programme component |
---|
Objectives of Programme's Learning Outcomes
Learning Outcomes of UE
This class serves as an introduction to first-order methods to tackle large-scale ooptimization problems that arise for example in data analysis and machine learning. A particular attention will be given to stochastic gradient methods that are used to compute the weights in deep neural networks.
UE Content: description and pedagogical relevance
Organization of the calss: - Introduction and motivation to the use of firt-order methods. - Optimal first-order methods for convex optimization. - Stochastic gradient methods. - Project: Comparison of first-order methods for solving a classification problem.
Prior Experience
Non-linear optimization
Type of Teaching Activity/Activities
AA | Type of Teaching Activity/Activities |
---|---|
I-MARO-303 |
|
I-MARO-304 |
|
Mode of delivery
AA | Mode of delivery |
---|---|
I-MARO-303 |
|
I-MARO-304 |
|
Required Learning Resources/Tools
AA | Required Learning Resources/Tools |
---|---|
I-MARO-303 | Slides and other references available on Moodle |
I-MARO-304 | Slides and other references available on Moodle |
Recommended Learning Resources/Tools
AA | Recommended Learning Resources/Tools |
---|---|
I-MARO-303 | Bottou, L., Curtis, F. E., & Nocedal, J. (2018). Optimization methods for large-scale machine learning. Siam Review, 60(2), 223-311. Newton, D., Yousefian, F., & Pasupathy, R. (2018). Stochastic Gradient Descent: Recent Trends. In Recent Advances in Optimization and Modeling of Contemporary Problems (pp. 193-220). INFORMS. |
I-MARO-304 | Bottou, L., Curtis, F. E., & Nocedal, J. (2018). Optimization methods for large-scale machine learning. Siam Review, 60(2), 223-311. Newton, D., Yousefian, F., & Pasupathy, R. (2018). Stochastic Gradient Descent: Recent Trends. In Recent Advances in Optimization and Modeling of Contemporary Problems (pp. 193-220). INFORMS. |
Other Recommended Reading
AA | Other Recommended Reading |
---|---|
I-MARO-303 | Not applicable |
I-MARO-304 | Not applicable |
Grade Deferrals of AAs from one year to the next
AA | Grade Deferrals of AAs from one year to the next |
---|---|
I-MARO-303 | Unauthorized |
I-MARO-304 | Authorized |
Term 2 Assessment - type
AA | Type(s) and mode(s) of Q2 assessment |
---|---|
I-MARO-303 |
|
I-MARO-304 |
|
Term 2 Assessment - comments
AA | Term 2 Assessment - comments |
---|---|
I-MARO-303 | This AA is evaluated via the project |
I-MARO-304 | This AA is evaluated via the project |
Term 3 Assessment - type
AA | Type(s) and mode(s) of Q3 assessment |
---|---|
I-MARO-303 |
|
I-MARO-304 |
|
Term 3 Assessment - comments
AA | Term 3 Assessment - comments |
---|---|
I-MARO-303 | idem Q1 |
I-MARO-304 | idem Q1 |