Great Ideas in Learning & Control Theory (5 credits)
Period 2 2023
Description
Certain ideas in learning and control theory have had a major impact on engineering, science and society at large during the 20th century. In this course we aim to introduce them to PhD students who may come from variety of different disciplinary backgrounds.
By taking this course, the student should be able to
- explain foundational concepts within statistical learning and control theory;
- implement, in cooperation, a solution to a cross-disciplinary problem;
- concisely summarize key ideas conveyed in the lectures below.
Course schedule
| Lec. | Date | Room | Topics |
|---|---|---|---|
| L1 | Fri Oct 20nd, 13-15 | Å 101127 | Entropy and information |
| L2 | Thu Oct 26th, 13-15 | Å 101127 | Dynamical system and state-space |
| L3 | Fri Nov 3rd, 13-15 | Å 101142 | Frequency domain |
| L4 | Thu Nov 9th, 13-15 | Å 101142 | Maximum likelihood and risk minimization |
| L5 | (postponed) | Å 101142 | Feedback and instability |
| L6 | Fri Nov 24th, 13-15 | Å 101142 | Stochastic approximation and adaptation |
| L7 | Fri Dec 1st, 13-15 | Å 101127 | Principle of optimality |
| L8 | Thu Dec 7th, 13-15 | Å 101127 | Convex optimization |
Map
over Ångström buildings
Examination
- Attendance of 6 out of 8 lectures is required.
- Preparation of a summary of the previous lecture (elevator pitch, ~5 min)
- Small collaborative project work
Prerequisites
Undergraduate courses in linear algebra and probability theory.
Registration
Send an e-mail to Per Mattsson
Teachers
Dave Zachariah, André Teixeira, Sergio Pequito and Per Mattsson