FJL3380 Theoretical Foundations of Machine Learning 6.0 credits

Content and learning outcomes
Course contents
A preliminary structure is given below:
Lecture 1. Introduction
Lecture 2. Probably Approximately Correct Framework and Empirical Risk Minimization
Lecture 3. Concentration inequalities
Lecture 4. The Vapnik-Chervonenkis (VC) Theory
Lecture 5. Linear Classification and Regression
Lecture 6. Regularization, Stability and Optimization
Lecture 7. Support Vector Machines and Kernel Methods
Lecture 8. Deep neural networks
Lecture 9. Clustering. Cluster validation and algorithms
Lecture 10. Reinforcement learning: model-free vs model-based algorithms
Lecture 11. Reinforcement learning: function approximation and deep RL
Intended learning outcomes
When completing the course, the student should be able to:
• Derive and apply the essential theoretical tools used in modern machine learning
- Concentration of measure in probability theory
- Stochastic optimization methods
- VC theory
• Describe the historical development of supervised and unsupervised learning algorithms
• Reflect on the advantages and drawbacks of deep learning
• Describe and explain the basic reinforcement learning algorithms and their modern versions
Course disposition
Lectures on selected topics.
Literature and preparations
Specific prerequisites
Recommended prerequisites
Equipment
N/A
Literature
Understanding Machine Leanring: From theory to algorithms, Shalev-Shwartz et al., lecture slides
Examination and completion
If the course is discontinued, students may request to be examined during the following two academic years.
Grading scale
Examination
- EXA1 - Examination, 6.0 credits, grading scale: P, F
Based on recommendation from KTH’s coordinator for disabilities, the examiner will decide how to adapt an examination for students with documented disability.
The examiner may apply another examination format when re-examining individual students.
Other requirements for final grade
Passing Grade based on 72h home exam and final project. The project consists in reading a few recent papers published at relevant conferences (NIPS, ICML) on a selected topic (e.g. on theoretical justification of deep learning), and to write a state-of-the-art report on the topic including historical developments, recent results, and open problems (5 pages double column minimum).
Opportunity to complete the requirements via supplementary examination
Opportunity to raise an approved grade via renewed examination
Examiner
Ethical approach
- All members of a group are responsible for the group's work.
- In any assessment, every student shall honestly disclose any help received and sources used.
- In an oral assessment, every student shall be able to present and answer questions about the entire assignment and solution.
Further information
Course web
Further information about the course can be found on the Course web at the link below. Information on the Course web will later be moved to this site.
Course web FJL3380