Abstract:
Recent research in diagnosis (Mcllraith, 1997; Thielscher, 1997; McIlraith, 1998) extends earlier works in diagnosis fi'om first principles by using an action theory instead of a first-order theory to describe the correct behavior of the system in consideration. The action theory allows us to reason about actions and their effects. Thus, when an action does not yield the expected effects something did mMfunction. Using regression, we can identify what is wrong initially. Our position is that in many cases this is not enough.