IPLM: A Medical Decision-Making System Based on Planning and Learning Methods

Contents

Medical diagnosis and therapy pl~nnlng can be viewed as a planning task with a incremental learning process, which is one of the important reason why medical tasks are dii~]cnit to implement on computers naturally. In this paper, we propose integration of reasoning, planning and learning methods in order to solve this dililcnity. The main point is that when medical experts deal with complicated cases, they apply integration of planning and learning methods to such diIBcult examples. Based on these concepts, we are now developing a system, IPLM (Interaction between Planning and Learning Methods). The resultshow that this methodology implements reasoning of medical experts more naturally, but also show that sometimes medical reasoning is a complicated mixture of diagnosis and treatment.