We can sometimes enforce abductive reasoning by using deduction. For instance, if we only have the rule rain → wetgrass, the only method for wetgrass is rain, and we can complete this implication with the opposite one wetgrass → rain.
With the three implications we had in the video example, the completion would be instead the more complex formula:
wetgrass → rain ∨ (sprinkle ∧ night) ∨ (glass ∧ fill ∧ push)
This method of completion can always be applied on normal logic programs. In fact, stable models are classical models of completion (but not vice versa). The completion method was first proposed by Keith C**** in 1978. Which is the surname of this author? (use capital letters)
