Fuzzy logic is the logic where every statement can have any real truth value between 0 and 1.

How can fuzzy logic be used in creating AI? Is it useful for certain decision problems involving multiple inputs? Can you give an example of an AI that uses it?


2 Answers 2


A classical example of fuzzy logic in an AI is the expert system Mycin.

Fuzzy logic can be used to deal with probabilities and uncertainties.

If one looks at, for example, predicate logic, then every statement is either true or false. In reality, we don't have this mathematical certainty.

For example, let's say a physician (or expert system) sees a symptom that can be attributed to a few different diseases (say A, B and C). The physician will now attribute a higher likelihood to the possibility of the patient having any of these three diseases. There is no definite true or false statement, but there is a change of weights. This can be reflected in fuzzy logic, but not so easily in symbolic logic.


My impression is that fuzzy logic has mostly declined in relevance and probabilistic logic has taken over its niche. (See the comparison on Wikipedia.) The two are somewhat deeply related, and so it's mostly a change in perspective and language.

That is, fuzzy logic mostly applies to labels which have uncertain ranges. An object that's cool but not too cool could be described as either cold or warm, and fuzzy logic handles this by assigning some fractional truth value to the 'cold' and 'warm' labels and no truth to the 'hot' label.

Probabilistic logic focuses more on the probability of some fact given some observations, and is deeply focused on the uncertainty of observations. When we look at an email, we track our belief that the email is "spam" and shouldn't be shown to the user with some number, and adjust that number as we see evidence for and against it being spam.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .