+Inference

Boolean inference

At its most basic level, inference is concerned with determining the truth value of a statement.

Thus, the following questions become important: What logic/ calculus is expressive enough, models the world naturally? How to reason about the the world? Is the logic/ calculus complete (ie: can we prove any true statement)?

Satisfiability vs validity

Given some axioms, can you prove a statement? When is a proposition with free variables satisfiable? Is it valid for all assignments?

Model checking

Often, we are given a knowledge-base, a set of statements \(S1\) known to be true in a certain model. Then, there are statements \(S2\) whose truth value is to be determined. Thus, we are to check whether the model given by \(S1 \union S2\) is consistent.

Probabilistic inference

More generally, it is concerned with evaluation of probabilities of events. This is accomplished using the axioms of probability, with much theory and devices like graphical models. See statistics and randomized algorithms surveys.