COMP312P - Machine Learning and Neural Computing
This database contains the 2017-18 versions of syllabuses. Syllabuses from the 2016-17 session are available here.
Note: Whilst every effort is made to keep the syllabus and assessment records correct, the precise details must be checked with the lecturer(s).
This module will introduce students with limited mathematics or computer science background to the fundamental concepts of machine learning and neural computing and their relation to intelligent systems. Machine learning is an exciting and important topic that has application both within and beyond the field of intelligent systems.
This module has two main learning outcomes:
1. Develop a knowledge and an understanding of machine learning and neural computing and how they relate to intelligent systems.
2. Gain experience in building basic machine learning and neural computing systems whilst understanding the limitations of such systems (e.g. over-fitting, curse of dimensionality, etc.).
The proposed course outline is as follows:
Topic 1: Introduction to machine learning
Introducing fundamental concepts and classes of machine learning (i.e. unsupervised. reinforced, supervised).
Topic 2: Classification & Regression
The two primary tasks of machine learning. Present the concepts of automated reasoning and decision making from learnt data models.
Topic 3: The neural model
Introduce neural computing as an alternative knowledge acquisition/representation paradigm along with its relationship to neurobiological models.
Topic 4: Deep representations
Introduce the concepts behind deep learning and the benefits of deep over shallow networks.
Topic 5: Reinforcement Learning
Introduce concepts behind learning the actions and behaviours that systems can undertake.
Topic 6: Effective use of machine learning
Conditions under which machine learning will / will not be successful. Also includes discussion of issues such as training and cross-validation.
Method of Instruction
The module will attempt to balance both the theory and practice of machine learning and neural computation for their use in intelligent systems.
Each week consists of two hours of lectures on theory and two hours of laboratories in which students will investigate the behaviour of those theories.
Assessment will be through completing a set of three or tour mini-projects will be set as part of the laboratory exercise. These projects will be assessed by demonstration of results and submission of a written report.
The course has the following assessment components:
Lab based worksheets (40%)
Reading list available via the UCL Library catalogue.