Reasoning by plausible inference

The PI (Plausible Inference) system has been implemented with the ability to reason in both directions, unlike previous AI systems for inexact reasoning. This is combined with truth-maintenance, dependency-directed backtracking, and time-varying contexts to permit modelling dynamic situations. Four rules of inference are employed in PI, Modus Ponens, Modus Tollens, Confirmation, and Denial. Each of these causes credibility as well as truth value to be propagated in a semantic network of assertions. Applications are foreseen in ‘guessing’ the answers to problems in formal deduction so that search trees may be pruned, and in the flexible selection of strategies for planning, as well as diagnostic or trouble-shooting models involving less-than-certain assertions.