Intelligent knowledge based systems (IKBS) are the central elements of fifth generation computers. They use inference to apply knowledge to perform a task. They require at least the following capabilities: classification, concept formation, summarizing, selection, searching, reasoning, planning, modelling, the use of ‘common sense' rules, and the ability to learn. Most of these are beyond the capabilities of present-day computers.
Research and development work is under way on various aspects of these problems, the most important - and the most difficult - being аn adequate method of knowledge representation. A number of techniques of knowledge processing are being tried out. These include evidential reasoning, based on the way people draw conclusions from bodies of evidence, such as that presented as a trial, and procedural learning, based on the way experts learn to solve difficult problems. Efficient methods of searching large databases are also being investigated.
Intelligent user interfaces
Intelligent user interfaces aim to make fifth generation computers usable by far more people than use computers at present. This is in spite of the complexity of the systems, and the fact that many users will not be computer experts. The technique is to make the computer behave much more as a person would expect. Two main channels of communication are being emphasized: the use of natural language and image processing.
Although a complete natural language will be beyond the capabilities of even a fifth generation computer, the intention is to enable them to use a large enough subset to make a number of voice-driven applications possible. Image processing requires the computer to be able to accept video pictures as input, and generate output of the same quality. Some approximation to three-dimensional vision is required. Intelligent interfaces will be supported by a complete hardware and software subsystem of the computer, as described above.
Applications of fifth generation computers
A very wide range of applications are planned for fifth generation computers, although their precise capabilities will not become clear until more development work has been done. Many applications will be expert systems of some sort.
The main industrial application is likely to be intelligent robots, with some degree of visual perception, and the ability to be instructed to attain certain goals, rather than to perform a sequence of steps, as at present. The robot will work out the sequence of steps to be performed, according to the goals, and external conditions of which it is aware.
Military applications include strategic and tactical planning and decision support systems, and automated weapons systems. The latter include intelligent guidance systems for missiles, and missile defence systems for aircraft and ships. Digital communications systems which are secure against electronic counter-measures are a likely by-product of VLSI chip design.
In commerce, corporate knowledge bases are likely to be constructed on top of corporate databases. Decision support systems for managers will use these knowledge bases to assist with planning and evaluation. The voice-activated word processor is the ultimate aim of a number of fifth generation projects, with a very large potential market awaiting the first team to achieve it. Automatic, or (more realistically) semi-automatic language translation systems are a similar application.
Conclusion
The fifth generation development projects are the main thrust of computer development for the next five to ten years. Large sums of money, and the time and energy of some of the leading IT experts in the world are being devoted to the work. The risks are very high, since many of the principles on which fifth generation computers are to be built did not exist five years before the start of the projects. However, the risks of withdrawing from the work are even higher. It is quite clear that if any national or regional group achieves a significant lead in the development of fifth generation computers, that group will be the dominant force in information technology.
