Inconsistency in definitions

Master / Vertiefung Künstliche Intelligenz
1

Disclaimer: Dieser Thread wurde aus dem alten Forum importiert. Daher werden eventuell nicht alle Formatierungen richtig angezeigt. Der ursprüngliche Thread beginnt im zweiten Post dieses Threads.

2

Inconsistency in definitions
On slide 232 in the notes we defined C:={C_uv| u,v ∈ V, u≠v}. How can on slide 250 C_uv not be in C? I think you mean C_uv=D_u x D_v instead of C_uv ∉C

3

But Cᵤᵥ = Dᵤ x Dᵥ would imply C’ᵤᵥ ⊆Cᵤᵥ so that case distinction wouldn’t make sense anymore…

My guess is that slide 250 is based on a slightly different definition where Cᵤᵥ ∉ C is a different way of saying Cᵤᵥ = Dᵤ x Dᵥ, i.e. there are no constraints between U and V. From a mathematical point of view it doesn’t help us but when it comes to implementations it could really help when the constraints are “sparse”.

4

My guess is that it shoud be C’[sub]uv[/sub] ∉ C that would mean that we have added a constraint that isn’t in the original C.

1 „Gefällt mir“
5

Thanks for the notification, that is actually a typo, it should be C’_uv\notin C’, i.e. the C’_uv is new.
The idea is that there are two ways of tightening: adding constraints and making the relations C_uv smaller.
I have fixed this in the slides.

6

I think you misunderstood my problem.

C’_uv ∉ C’ can not be true according to definition on slide 232 in my opinion.

Thats exactly the point. I think you can just cut the first part and everything is fine and exactly as Prof. Kohlhase wanted it to be :).