working with examples

I have made these notes
<https://docs.google.com/document/d/1lYn8-YUvIS_k1rezgqdR0DX2BLTRYpvkJpmvYjAEpng/edit?usp=sharing>
editable, however, the wiki and other pages on the AI KR CG have always
been more or less editable
Dave, you can also add new tentative goals to the CG, to be discussed
if the proposed ones are not defined enough

You brought up a couple of good examples recently the stable diffusion and
the ai for deep fakes, can we learn something from them?
does anyone have other examples of ML that may or may not use KR
and can we use them as working examples to understand how  KR can work for
ML
Please add more working examples to the table in doc and lets discuss them
some time. I am very short of eyesight lately, apologies if I do not
read/reply to posts,

CASE

USE OF KR?

LIMITATION/PROBLEM

POSSIBLE KR SOLUTION?/IMPROVEMENT

Other points to be discussed?

STABLE DIFFUSION

NO

CANNOT TELL WHAT IT DOES AND HOW IT DOES IT

DEVISE KR TO MAKE EXPLICIT THE LOGIC

AI FOR DEEP FAKES

YES

?


…..