GEORGIA INST OF TECH ATLANTA SCHOOL OFINFORMATION AND COMPUTER SCIENCE
The objective of this research was to elucidate the role of experience in common-sense and expert problem solving. The aim was to discover and describe the processes involved in extracting useful conceptual knowledge from experience, in organizing and building the schemata to hold that knowledge, and in using that information in problem solving. In research areas as diverse as natural language processing and expert systems, researchers are plagued by the fact that the knowledge the systems need is hard to collect and input to the system. One way this bottleneck, called knowledge acquisition, can be relieved is by providing systems with a means of learning from their experiences. This research helps to lay the theoretical foundation for reasoning systems that 1 can become more expert through experience, 2 can make predictions and give advice based on previous experience in similar situations, and 3 can adapt to changes in their environments.