fuzzy logic

Terms from Artificial Intelligence: humans at the heart of algorithms

The glossary is being gradually proof checked, but currently has many typos and misspellings.

Fuzzy logics are one of the forms of fuzzy reasoning. They assign levels of 'truth' to varables rater than the dichotomous true/false of Boolean logic. This is used to capture the levels of uncertainty we have in real life about many things. So, for example, we might have:

    is_summer=1 – definitely true

    feel_like_break=0.7 – more true than not

    will_stay_fine=0.2 – closer to false (UK weather!)

Logical operators AND and OR are converted into max and min , which correspond to the normal logic meanings for true/false values. We can then create a rule such as:

    go_for_a_picnic   =   is_summer AND ( feel_like_break OR will_stay_fine

        =   min ( 1, max( 0.7, 0.2 ) )   =   min(1,0.7) = 0.7

Used in Chap. 3: pages 30, 34

Also known as fuzzy