F20BC - Biologically Inspired Computation

Course leader(s):

Michael Adam Lones (Edinburgh, September)
Wei Pang (Edinburgh, September)
Hadj Batatia (Dubai, September)
Claudio Zito (Dubai, September)

Aims

Traditional computation finds it either difficult or impossible to perform a certain key range of tasks associated with pattern recognition, problem solving and autonomous intelligence. Great progress towards designing software for such tasks has emerged by taking inspiration from a range of natural, mainly biological, systems.

The aims of this course are to:

introduce an appreciation of the former
introduce the main biologically-inspired algorithms and techniques which are now commonly researched and applied
establish a practical understanding of the real-world problems to which these techniques may be fruitfully be applied.

Syllabus

1. Neurocomputing and deep learning (1.1 1. Gradient descent, 1.2 2. Neural networks, 1.3 3. Deep learning, 1.4 4. Neuroevolution)

2. Evolutionary computation (2.1 1. Optimisation, 2.2 2. Evolutionary algorithms, 2.3 3. Genetic programming, 2.4 4. Multiobjective evolutionary algorithms)

3. Swarm computing (3.1 1. Swarm intelligence, 3.2 2. Ant colony optimisation, 3.3 3. Particle swarm optimisation)

4. Cellular automata (4.1 1. Computational universality, 4.2 2. Game of life, 4.3 3. Elementary cellular automata)

Learning outcomes

By the end of the course, students should be able to do the following:

explain limitations of traditional computation models
analyse a range of biologically inspired computation models and algorithms, explain their limitations, and illustrate areas of applicability
implement one or more biologically inspired models and algorithms to solve a practical problem
identify bio-inspired models and algorithms that apply to optimisation and machine learning problems
exercise substantial autonomy and initiative when solving a real-world problem using bio-inspired techniques
demonstrate critical reflection when analysing experimental results and comparing bio-inspired models and algorithms

Further details

Curriculum explorer: Click here

SCQF Level: 10

Credits: 15