Inverse approximate reasoning: maximize entropy

We consider the problem of given the conclusion of a fuzzy rule, find in some sense the "best" data (input) that will produce the given conclusion. In this paper we investigate the use of the principle of maximum entropy to solve this problem.