Lec 12 不确定性推理 I1

Reasoning under uncertainty

Review of probabilities, independence, and Bayes rule

Bayes net: concept, construction, inference, and independence

Variable elimination algorithm

*Slides based on those of Sheila McIlraith

Y. Liu Intro to AI 1 / 62

Uncertainty

In search we viewed actions as being deterministic.

executing action A in state S

1

causes transition to state S

2

Y. Liu Intro to AI 2 / 62

Uncertainty

We might not know exactly what state we start off in

e.g., we can’t see our opponents’ cards in a poker game

We don’t know what a patient’s ailment is.

We might not know all of the effects of an action

The action might have a random component, like rolling dice.

We might not know all of the long term effects of a drug.

An action might fail

Y. Liu Intro to AI 3 / 62

Uncertainty

In such domains we still need to act,

but we can’t act solely on the basis of known true facts.

We have to /gamble0.

Y. Liu Intro to AI 4 / 62

An example

is low.

If we must arrive at the airport at 4:30pm on Friday

we most likely need 1.5 hour or more to get to the airport.

Y. Liu Intro to AI 5 / 62