2
1) Recall from Q-learning that Q(a,i) represents the expected utility of doing action a in state i. Recall also that these expected
values are updated during processing. The updating rule looks something like:
Q (a, i ) = Q( a, i ) +
( new )
R(i ) + min
Name _ 2
1. (14 points) We estimate the accuracy of a particular hypothesis h from its performance
on a conventional training set. After 5 examples, the accuracy of h is 0.8
Probability Density
As discussed in class, the sample average of 0.8, based on ju
Name_
1. (30 pts.) Show that you understand the following terms by giving two important facts about each. You might consider
such facts as a definition, an illustration, a comparison with some other relevant concept, or an important property. Use no
more
CS440: Introduction to AI
Homework 5
Due: Friday, March 22th, 2013
1
Basic Probability Review (30 points)
1. A biased die has the following probabilities of landing on each face:
face
123456
P (face) .2 0 .1 .3 .3 .1
What is the probability that the die s
CS 440 / ECE 448: Introduction to AI
Homework 3
1. Planning Actions and States
(a) P eg (P 0) P eg (P 1) P eg (P 2) Disk (D0) Disk (D1) Disk (D2) T able(T bl)
Empty (P 1) Empty (P 2)
Bigger(D0, D1) Bigger(D1, D2) Bigger(D0, D2)
On(D0, P 0) On(D1, P 0)
CS440: Introduction to AI
Homework 2: Solutions
Due: Wednesday, February 13th, 2013
1 First Order Predicate Calculus
This problem has two parts. In the rst part, you are asked to translate English sentences into rst order
predicate calculus (FOPC) sentenc
CS 440 / ECE 448: Introduction to AI
HW 1 Solutions
Problem 1. Suppose you want to implement an Android app to aid a
mail carrier (with access to Google Maps, Google Earth, etc). Consider the
following alternative models to estimate the time it would take