A general numerical approach to the benchmark problems in defeasible reasoning

Researchers in the area of defeasible reasoning have assembled a collection of benchmark problems which typify situations where such reasoning seems called for. The author applies the method of path analysis to a number of these problems. He shows how this statistical technique can be re-interpreted as a probabilistically correct method for belief propagation in reasoning systems. His intent is to illustrate this method through examples, and to show that the benchmark problems can be capably handled quantitatively