This is the first year funding of a three-year continuing award to study uncertain inference. While there may be a variety of sources and kinds of uncertainty, the kind of uncertainty that is most pervasive and that should be best understood is uncertainty based on (at least approximate) statistical knowledge. This project focuses on that kind of uncertainty. There are three main applications of statistical knowledge in planning under uncertainty; in justifying the assumptions we take for granted in planning and in design, and in obtaining and updating statistical knowledge. When it comes to making these applications under the special, unique, never-to-be repeated circumstances in which we find ourselves, they are each problematic. In each case the key concept is that of uncertain inference. We will characterize uncertain inference, and develop principles that will allow the formal application of statistical knowledge in the presence of rich and detailed knowledge of our unique circumstances.
