機器學習理論
Theoretical Machine Learning
| 節 | 週二 |
|---|---|
5 13:20–14:10 | 機器學習理論 EDB07 3 節連堂 |
6 14:20–15:10 | |
7 15:30–16:20 |
* 根據陽明交大上課時間表所列
This is an advanced graduate course in the theory of machine learning. The course is ideal for graduate students and senior undergraduates who are theoretically inclined and want to know more about related research challenges in the field of machine learning. A tentative list of topics include the following:
Probability, Linear Algebra, Optimization
Like all my courses, the class will have 2hrs of in-class lecture and 1hr of quiz every week. The quiz will alternate between theory and programming quizzes. The class will also have a midterm and a final. A project will also be developed throughout the course. Evaluation and Grading Policy: Project, quizzes and exams will all account for one third of the grade. Pedagogy and other supplementary information (websites, TAs, handouts and/or databases): I will develop slides for this course and post the project online on my personal website
| 週次 | 主題 |
|---|---|
| 第 1 週 | Introduction, learning theory and statistical learning theory |
| 第 2 週 | VC theory |
| 第 3 週 | PAC model |
| 第 4 週 | Metric spaces, Convex sets, supporting hyperplanes, |
| 第 5 週 | Duality theory, Lagrangian |
| 第 6 週 | Primal & Dual problem, Strong convexity, Lipschitz continuity |
| 第 7 週 | Gradient descent |
| 第 8 週 | Subgradient descent, Generalized gradient descent |
| 第 9 週 | Acceleration methods |
| 第 10 週 | Matrix differentiation and matrix algebra |
| 第 11 週 | Newton's method |
| 第 12 週 | Linear Programming |
| 第 13 週 | Duality theory and KKT conditions |
| 第 14 週 | Alternating Direction Method of Multipliers |
| 第 15 週 | Support Vector Machines |
| 第 16 週 | Interior point Methods |
| 第 17 週 | Non-convexity theory |
| 第 18 週 | Convex approaches to solving non convex functions |
Understanding Machine Learning: From Theory to Algorithms Shai Ben-David and Shai Shalev-Shwartz