The best of two artificial intelligence forms
Researchers at The Chinese University of Hong Kong have been looking at how to get the best from two different forms of artificial intelligence.
On one side are conventional expert computer systems which mimic human decision-making and solve problems using if ...then rules. They are capable of commonsense reasoning with everyday language. On the other side are neural networks which model human brain mechanisms and can identify and extract patterns and trends from sounds and other data sensed externally.
While expert systems are efficient at the symbolic representation of knowledge but are inefficient at automatic knowledge acquisition and learning, neural networks have a strong capacity to learn but symbol processing is more complex.
Researchers at CUHK have been developing a way of making both systems complement each other by combining them into a hybrid for use in solving real-life decision-making problems.
Examples of application are the decision-making support systems needed for computerised environment assessment and management as, for example, in water pollution management, and the system support needed for computerised medical diagnosis.
Principal Investigator,
Prof Yee Leung, Department of Geography,
email: yeeleung@cuhk.edu.hk |