Back to Study Material Artificial Intelligence

Fuzzy logic

Classical logic operates on crisp binary truth values: a proposition is either completely true or completely false. This works well for formal mathematics ...

Fuzzy logic

Beyond True and False

Classical logic operates on crisp binary truth values: a proposition is either completely true or completely false. This works well for formal mathematics but poorly captures many concepts in human reasoning. Is a person who is one hundred and eighty centimetres tall "tall"? The honest answer is "somewhat," not a definitive yes or no. Fuzzy logic, introduced by Lotfi Zadeh in 1965, extends classical logic to handle this inherent vagueness by allowing truth values to range continuously between zero and one.

In fuzzy logic, membership functions map crisp input values to degrees of membership in fuzzy sets. A tall membership function might assign a value of zero to heights below one hundred and sixty centimetres, rise gradually to one at heights above one hundred and ninety centimetres, and take intermediate values between. A person at one hundred and seventy five centimetres might have a tallness membership of zero point six, meaning they are sixty percent tall by this definition.

Fuzzy inference systems apply fuzzy rules to derive fuzzy outputs, which are then converted back to crisp values through defuzzification. The most common defuzzification method is centroid defuzzification, which computes the centre of mass of the resulting fuzzy output region. A washing machine using fuzzy logic might have rules like "if the water is very dirty and the load is heavy then wash for a long time," where "very dirty" and "heavy" and "long time" are all fuzzy sets with membership functions rather than hard thresholds.

Fuzzy vs Probability

A common point of confusion is the difference between fuzzy membership and probability. Probability expresses uncertainty about whether something is true. Fuzzy membership expresses the degree to which something belongs to a category. A glass half full of water has a fuzzy fullness membership of zero point five, but there is no uncertainty; we know exactly how full it is.

Interview Tip

Interviews may ask where fuzzy logic is appropriate versus probabilistic methods. Fuzzy logic excels at encoding linguistic human knowledge into control systems, such as climate control and anti-lock braking systems, where engineers can express rules in natural language terms. Probabilistic methods are preferred when there is genuine uncertainty that can be quantified from data. Mentioning automotive control systems as a concrete application domain is effective.