Close
Description
Aims:
The module aims to introduce students to the foundations of reinforcement learning, and to equip students with the ability to successfully implement, apply and test relevant learning algorithms.
Intended learning outcomes:
On successful completion of the module, a student will be able to:
- Understand the basics of reinforcement learning paradigms.
- Understand the theoretical foundations, formalisms and algorithms in reinforcement learning.
- Understand how to apply reinforcement learning algorithms to environments with complex dynamics.
- Understand how to combine reinforcement learning with function approximation, and specifically with modern deep learning methods (deep reinforcement learning).
Indicative content:
The module is about prediction and control using reinforcement learning, including aspects of deep reinforcement learning, i.e., the application of neural networks-based functional approximation to reinforcement learning problems.
The following are indicative of the topics the module will typically cover:
- Markov decision processes.
- Planning by dynamic programming.
- Model-free prediction and control.
- Value function approximation.
- Policy gradient methods, Actor-critic algorithms.
- Integration of Learning and Planning.
- Exploration vs exploitation trade-offs.
For these topics we will discuss theory and concrete algorithms and applications.
Requisites:
To be eligible to select this module as an optional or elective, a student must: (1) be registered on a programme and year of study for which it is formally available; (2) have a strong understanding of probability, calculus, and linear algebra; (3) have knowledge of coding skills in Python (in order to complete assessments); and (4) have taken at least one introductory machine learning module, for example Supervised Learning (COMP0078) or Introduction to Machine Learning (COMP0088) (or be concurrently enrolled in such a module).
It is recommended that students have taken either Bayesian Deep Learning (COMP0171), or Applied Deep Learning (COMP0197) (or be concurrently enrolled in such a module).
Module deliveries for 2024/25 academic year
Last updated
This module description was last updated on 8th April 2024.
听