What you will study
With the advent of the digital computer in the twentieth century, genuine man-made intelligence seemed possible for the first time, and artificial intelligence (AI) emerged as a serious research discipline. In M366 you’ll be introduced to both conventional and novel ideas in AI, by contrasting traditional approaches with ideas that are now being pursued in the latest research– taking in aspects of biology and philosophy as well as computing and technology. The module has three broad objectives, spread over six blocks.
First, it introduces and contrasts traditional and modern (sometimes called ‘nouvelle’) approaches to AI. Traditional AI research attempts to reproduce in computers some of the characteristics that we think of as central to human intelligence: logical reasoning, language, problem solving and our ability to plan and predict. The conventional strategy is to start by examining introspectively the workings of our own minds, and then try to replicate these on a computer. The computational techniques that come out of this approach are discussed in detail, with examples, in Blocks 1 and 2. In contrast, nouvelle AI research looks beyond the human sphere, to evidence of intelligence in non-human animals. These include self-organised collective behaviour; the ability to recognise objects and respond appropriately; communication; navigation; construction skills; and learning. Block 3 uses examples from nature and from computer simulation to develop the four main principles out of which such purposeful, systematic behaviours can arise: interaction, emergence, adaptation and selection.
Second, using the four principles as a base, M366 discusses in detail two modern techniques in AI and computerised problem solving: artificial neural networks (ANNs) and evolutionary computation (EC). The biological inspiration of ANNs in animal nervous systems is described in outline in Block 4, along with analysis and examples of successful ANN systems and models. EC is introduced in its natural context of genetic adaptation and Darwinian evolution in Block 5. Techniques such as genetic algorithms and genetic programming – in which evolutionary processes are simulated to solve problems of optimisation, control and design – are explained and analysed in detail.
Finally, in Block 6, students are invited to reflect on the contrasting traditional and modern approaches, and to form their own opinions on the significance and the future of AI.
The formal presentation of the above material will be backed up by experimental work using a number of software tools.
The module contains images of insects, including moving swarms in the video material on the DVD.