In computational learning theory there have been some interesting developments recently on the problem of ``learning from experts.''

In this paper, we ``adapt'' the learning problem to an adaptive control formulation. What results is an adaptive controller which is reminiscent of a certainty equivalence scheme using the ``posterior mean'' for the parameter estimator. We show that this scheme can be analyzed in a somewhat novel way, for ideal linear systems. The analysis techniques may be of some interest to researchers in the theory of adaptive control.

Back to Adaptive Control, Filtering, and Signal Processing Table of Contents