SOLUTIONS & INSTRUCTOR MANUAL
Artificial Intelligence: A Modern Approach, 4th Edition
by Peter Norvig and Stuart Russell, Chapters 1 – 28
,Artificial Intelligence yc
1 Introduction ...
ycycy cy c yc yc yc
2 Intelligent Agents ...
ycycy cy c yc yc yc yc
II Problem-solving
yc
3 Solving Problems by Searching ...
ycycy cy c yc yc yc yc yc yc
4 Search in Complex Environments ...
ycycy cy c yc yc yc yc yc yc
5 Adversarial Search and Games ...
ycycy cy c yc yc yc yc yc yc
6 Constraint Satisfaction Problems ...
ycycy cy c yc yc yc yc yc
III Knowledge, reasoning, and planning
yc yc yc yc
7 Logical Agents ...
ycycy cy c yc yc yc yc
8 First-Order Logic ...
ycycy cy c yc yc yc yc
9 Inference in First-Order Logic ...
ycycy cy c yc yc yc yc yc
10 Knowledge Representation ...
ycycy cy c yc yc yc yc
11 Automated Planning ...
ycycy cy c yc yc yc yc
IV Uncertain knowledge and reasoning
yc yc yc yc
12 Quantifying Uncertainty ...
ycycy cy c yc yc yc yc
13 Probabilistic Reasoning ...
ycycy cy c yc yc yc yc
14 Probabilistic Reasoning over Time ...
ycycy cy c yc yc yc yc yc yc
15 Probabilistic Programming ...
ycycy cy c yc yc yc yc
16 Making Simple Decisions ...
ycycy cy c yc yc yc yc yc
17 Making Complex Decisions ...
ycycy cy c yc yc yc yc yc
18 Multiagent Decision Making ...
ycycy cy c yc yc yc yc yc
V Machine Learning
yc yc
19 Learning from Examples ...
ycycy cy c yc yc yc yc yc
, 20 Learning Probabilistic Models ...
ycycy cy c yc yc yc yc yc
21 Deep Learning ...
ycycy cy c yc yc yc yc
22 Reinforcement Learning ...
ycycy cy c yc yc yc yc
VI Communicating, perceiving, and acting
yc yc yc yc
23 Natural Language Processing ...
ycycy cy c yc yc yc yc yc
24 Deep Learning for Natural Language Processing ...
ycycy cy c yc yc yc yc yc yc yc yc
25 Computer Vision ...
ycycy cy c yc yc yc yc
26 Robotics ...
ycycy cy c yc yc yc
VII Conclusions yc
27 Philosophy, Ethics, and Safety of AI ...
ycycy cy c yc yc yc yc yc yc yc yc
28 The Future of AI
ycycy cy c yc yc yc yc
, EXERCISES y c y c
1
INTRODUCTION
Note that for many of the questions in this chapter, we give references where answers can be fou
yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc
nd rather than writing them out—the full answers would be far too long.
yc yc yc yc yc yc yc yc yc yc yc yc
1.1 What Is AI?
y c yc yc
Exercise 1.1.#DEFA yc
Define in your own words: (a) intelligence, (b) artificial intelligence, (c) agent, (d) ra-
yc yc yc yc y c yc yc yc yc yc yc yc yc
tionality, (e) logical reasoning.
yc yc yc yc
a. Dictionary definitions of intelligence talk about “the capacity to acquire and apply kno yc yc yc yc yc yc yc yc yc yc yc yc
wledge” or “the faculty of thought and reason” or “the ability to comprehend and profit yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc
from experience.” These are all reasonable answers, but if we want something quantifi
yc y c yc yc yc yc yc yc yc yc yc yc
able we would use something like “the ability to act successfully across a wide range of ob
yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc
jectives in complex environments.” yc yc yc
b. We define artificial intelligence as the study and construction of agent programs that
yc yc yc yc yc yc yc yc yc yc yc yc
perform well in a given class of environments, for a given agent architecture; they do th
yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc
e right thing. An important part of that is dealing with the uncertainty of what the curre
yc yc y c yc yc yc yc yc yc yc yc yc yc yc yc yc
nt state is, what the outcome of possible actions might be, and what is it that we really d
yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc
esire.
c. We define an agent as an entity that takes action in response to percepts from an envi-
yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc
ronment. yc
d. We define rationality as the property of a system which does the “right thing” given w
yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc
hat it knows. See Section 2.2 for a more complete discussion. The basic concept is perf
yc yc y c yc yc yc yc yc yc yc y c yc yc yc yc
ect rationality; Section ?? describes the impossibility of achieving perfect rational-
yc yc yc yc yc yc yc yc yc yc
ity and proposes an alternative definition.
yc yc yc yc yc yc
e. We define logical reasoning as the a process of deriving new sentences from old, such th
yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc
at the new sentences are necessarily true if the old onesare true. (Notice thatdoes not refer
yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc
to anyspecific syntax or formallanguage, but it doesrequire a well-
yc yc yc yc yc yc yc yc yc yc yc yc
definednotion of truth.) yc yc yc
Exercise 1.1.#TURI yc
Read Turing’s original paper on AI (Turing, 1950). In the paper, he discusses several objecti
yc yc yc yc yc yc yc y c yc yc yc yc yc yc
ons to his proposed enterprise and his test for intelligence. Which objections still carry
yc yc yc yc yc yc yc yc yc yc yc yc yc
© 2023 Pearson Education, Hoboken, NJ. All rights reserved.
yc yc yc yc yc yc yc yc
Artificial Intelligence: A Modern Approach, 4th Edition
by Peter Norvig and Stuart Russell, Chapters 1 – 28
,Artificial Intelligence yc
1 Introduction ...
ycycy cy c yc yc yc
2 Intelligent Agents ...
ycycy cy c yc yc yc yc
II Problem-solving
yc
3 Solving Problems by Searching ...
ycycy cy c yc yc yc yc yc yc
4 Search in Complex Environments ...
ycycy cy c yc yc yc yc yc yc
5 Adversarial Search and Games ...
ycycy cy c yc yc yc yc yc yc
6 Constraint Satisfaction Problems ...
ycycy cy c yc yc yc yc yc
III Knowledge, reasoning, and planning
yc yc yc yc
7 Logical Agents ...
ycycy cy c yc yc yc yc
8 First-Order Logic ...
ycycy cy c yc yc yc yc
9 Inference in First-Order Logic ...
ycycy cy c yc yc yc yc yc
10 Knowledge Representation ...
ycycy cy c yc yc yc yc
11 Automated Planning ...
ycycy cy c yc yc yc yc
IV Uncertain knowledge and reasoning
yc yc yc yc
12 Quantifying Uncertainty ...
ycycy cy c yc yc yc yc
13 Probabilistic Reasoning ...
ycycy cy c yc yc yc yc
14 Probabilistic Reasoning over Time ...
ycycy cy c yc yc yc yc yc yc
15 Probabilistic Programming ...
ycycy cy c yc yc yc yc
16 Making Simple Decisions ...
ycycy cy c yc yc yc yc yc
17 Making Complex Decisions ...
ycycy cy c yc yc yc yc yc
18 Multiagent Decision Making ...
ycycy cy c yc yc yc yc yc
V Machine Learning
yc yc
19 Learning from Examples ...
ycycy cy c yc yc yc yc yc
, 20 Learning Probabilistic Models ...
ycycy cy c yc yc yc yc yc
21 Deep Learning ...
ycycy cy c yc yc yc yc
22 Reinforcement Learning ...
ycycy cy c yc yc yc yc
VI Communicating, perceiving, and acting
yc yc yc yc
23 Natural Language Processing ...
ycycy cy c yc yc yc yc yc
24 Deep Learning for Natural Language Processing ...
ycycy cy c yc yc yc yc yc yc yc yc
25 Computer Vision ...
ycycy cy c yc yc yc yc
26 Robotics ...
ycycy cy c yc yc yc
VII Conclusions yc
27 Philosophy, Ethics, and Safety of AI ...
ycycy cy c yc yc yc yc yc yc yc yc
28 The Future of AI
ycycy cy c yc yc yc yc
, EXERCISES y c y c
1
INTRODUCTION
Note that for many of the questions in this chapter, we give references where answers can be fou
yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc
nd rather than writing them out—the full answers would be far too long.
yc yc yc yc yc yc yc yc yc yc yc yc
1.1 What Is AI?
y c yc yc
Exercise 1.1.#DEFA yc
Define in your own words: (a) intelligence, (b) artificial intelligence, (c) agent, (d) ra-
yc yc yc yc y c yc yc yc yc yc yc yc yc
tionality, (e) logical reasoning.
yc yc yc yc
a. Dictionary definitions of intelligence talk about “the capacity to acquire and apply kno yc yc yc yc yc yc yc yc yc yc yc yc
wledge” or “the faculty of thought and reason” or “the ability to comprehend and profit yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc
from experience.” These are all reasonable answers, but if we want something quantifi
yc y c yc yc yc yc yc yc yc yc yc yc
able we would use something like “the ability to act successfully across a wide range of ob
yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc
jectives in complex environments.” yc yc yc
b. We define artificial intelligence as the study and construction of agent programs that
yc yc yc yc yc yc yc yc yc yc yc yc
perform well in a given class of environments, for a given agent architecture; they do th
yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc
e right thing. An important part of that is dealing with the uncertainty of what the curre
yc yc y c yc yc yc yc yc yc yc yc yc yc yc yc yc
nt state is, what the outcome of possible actions might be, and what is it that we really d
yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc
esire.
c. We define an agent as an entity that takes action in response to percepts from an envi-
yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc
ronment. yc
d. We define rationality as the property of a system which does the “right thing” given w
yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc
hat it knows. See Section 2.2 for a more complete discussion. The basic concept is perf
yc yc y c yc yc yc yc yc yc yc y c yc yc yc yc
ect rationality; Section ?? describes the impossibility of achieving perfect rational-
yc yc yc yc yc yc yc yc yc yc
ity and proposes an alternative definition.
yc yc yc yc yc yc
e. We define logical reasoning as the a process of deriving new sentences from old, such th
yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc
at the new sentences are necessarily true if the old onesare true. (Notice thatdoes not refer
yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc yc
to anyspecific syntax or formallanguage, but it doesrequire a well-
yc yc yc yc yc yc yc yc yc yc yc yc
definednotion of truth.) yc yc yc
Exercise 1.1.#TURI yc
Read Turing’s original paper on AI (Turing, 1950). In the paper, he discusses several objecti
yc yc yc yc yc yc yc y c yc yc yc yc yc yc
ons to his proposed enterprise and his test for intelligence. Which objections still carry
yc yc yc yc yc yc yc yc yc yc yc yc yc
© 2023 Pearson Education, Hoboken, NJ. All rights reserved.
yc yc yc yc yc yc yc yc
