Uncertainty

• That truth thing from FOPL is pretty difficult to swallow for a lot of things.
• Fortunately, there are a lot of ways to deal with uncertainty.
• Probability: the most rigorous is probability. If the probability can be stated, then you can mathematically derive probabilities of consequents.
• The problem is that is really hard to do.
• Bayesian networks are a very successful extension of probability.
• Similarly, hidden markov models are related.
• Another question is about other people or agents knowledge states. This fits in with the Belief, Desires and Intentions work.
• A really popular way of dealing with uncertainty is Fuzzy logic. I think it's unprincipled, but it's quick and dirty.