0.9x0.9=0.81<0.9 and 1.1x1.1=1.21>1.1. Dont you feel this property distorts Deep Learning results by a little bit?

I know it seems trivial but I had to ask this. Don’t you think universal property of numbers is when two numbers multiplied among themselves, it creates results greater than each (outside 0-1 property) and this property changes when we are between 0 to 1. In deep Learning we multiply two numbers and add with other, that all we do. Here each numbers are not only numbers they are features too!! We defined MULTIPLICATION and ADDITION in mathematics and I feel this is not digesting news that suddenly propoerty changes when we are between 0 and 1 and when >1. Its like when two features are saying there is eye in any given image both with probability 0.8, we are making decision that its 0.64 (less confident) rather than interpreting in Random Forest terms like “Yipie we are glad to know that two are saying eyes are there in image with 80% probability.”. This last example is very loosely alinged in the direction of my question but I hope someone grasps my idea!

I just feel its the way we defined things in universe and its just not fitting all things together here!