The application of Artificial Intelligence (AI) to chemistry is in its early stages. Although knowledge-based systems have been used for some time in analytical chemistry, in most laboratories computers merely calculate - they do not take intelligent decisions.
Yet, as computer power grows, so does the potential of AI methods. We are exploiting this potential, using Genetic Algorithms, Neural Networks, Self-ordering Maps, Support Vector Machines and other techniques.
Such methods have been a part of AI for two decades or longer, but only recently has their potential within science been recognized. AI programs learn, instead of having to be told what to do, and are often suited to just those types of problems which give conventional programs the greatest difficulty.
Using Genetic Algorithms, we have studied the control of chemical reactors, the interpretation of AIDS-related data, the properties of magneto-rheological fluids, the dispersal of pollution and the rotational motion of proteins in solution. Work with Self-ordering Maps has included the assessment of geochemical data relating to hydrocarbon deposits, the analysis of NMR spectra from drug trials, and an investigation into the oxidation of wine.
At present, we are collaborating with an international oil company on the use of Artificial Intelligence in the new, and rapidly-developing, developing field of Petroleomics.
Part II students may work on AI projects, or as members of CoLoS, a science education consortium of 15 European universities. Our main interest in this area is in the development of on-line science experiments, accessible through the Internet. Part II projects within AI and the CoLoS scheme are suited to students with a strong interest in computing.