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.
Ass3: Stochastic Wumpus
Hi,
I was going through the exercises when I noticed something. It’s about Assignment 3.1: Stochastic Wumpus, specifcally 3.1.1.
In this subexercise we have to compute two things essentially:
- P(-W1|S1): no problem.
- P(W2|S1): this is what I dont understand. The solution tells me that P(S1|W2) = 0.2 which we need for the computation of P(W2|S1). What I understand is P(S1|-W1) = 0.2 , as it is also mentioned in the description. “Robby also thinks, that places wihtout a Wumpus should rarely stink (in only 20% of cases), whereas every field with a Wumpus stinks.” I read this as follows P(Si|-Wi) = 0.2 and P(Si|Wi) = 1. There is no word about whether the neighbourd fields stinks. The solution obviously assumed that P(Sj|Wi) = 0.2 for j ≠ i. What the solutions is probably trying to tell us is, that if we know the Wumpus is in i, and assuming there is only one Wumpus, it cannot be in Field j. Following this, we can conclude that P(Sj|Wi) = 0, because there is no Wumpus in j, right? There is also no other source of causing a stench, at least there is nothing mentioned about that.
So the question is, how can we conclude that P(Sj|Wi) = 0.2 for j ≠ i ?
Best
They can’t (any more than any other field without a wumpus does), since:
This already tells you the precise probability that a given field stinks given that a Wumpus is in the field, or given that there isn’t. Any modification regarding neighborhoods would violate this sentence.
Not quite. What we consequently can conclude is that P(Sj|Wi)=0.2 - because Wi implies -Wj and P(Sj|-Wj)=0.2