Great Ideas in Learning and Control
Seminar-style PhD course covering foundational topics at the intersection of learning and control theory.
Overview
This PhD seminar course (FTN0509) covers foundational ideas spanning learning and control theory. The course runs across two semesters with nine lectures followed by student presentations.
Topics
- Entropy and information
- Dynamical systems and state-space representations
- Frequency domain methods
- Maximum likelihood and risk minimisation
- Stochastic approximation and adaptation
- Principle of optimality
- Convex optimization
Format
Lecture-based sessions with paired discussion forums, culminating in final student presentations.