Logic for Computer Science and Artificial Intelligence

Chapter 11

Knowledge and Logic: Some Notions

In computer science and AI, people often talk of “knowledge”, “knowledge representation”, “efficient usage of knowledge”, etc.

We may wonder whether logic, which allowed us, for example, to verify whether a reasoning is correct, to obtain conclusions from premises, to reason with fuzzy concepts, to qualify formulas as necessary or possible and to deduce consequences, to take time into account in formulas, and to reason on these formulas…will also allow us to talk of (and define) knowledge and to reason while taking into account the knowledge of an agent in a world where there are other agents, possibly with a different knowledge.

The goal of this chapter is to try to answer this question.

Up to now, we have assumed (implicitly or explicitly) that “knowledge” was more or less a synonym of “premise” and that the initial premises did not evolve when they interacted with the environment.

But while handling problems in which the state of knowledge may change, new difficulties will arise.

In game theory (an entire discipline by itself that we only mention here), it is also necessary to take an environment that is not passive into account, meaning that the actions of the other agents in the environment are relevant in the design of strategies to reach a goal.

Recall that the ideas on probabilities play a very important role in knowledge theory (for the time being we admit the intuitive meaning of knowledge): it suffices to think of natural ...