chapter 13: artificial intelligence chapter 13 artificial intelligence page 139 computations that...

9
Chapter 13: Artificial Chapter 13: Artificial Intelligence Intelligence Chapter 13 Chapter 13 Artificial Artificial Intelligence Intelligence Page Page 1 Computations that Computations that make it possible for make it possible for a machine to a machine to perceive, reason, perceive, reason, and act in a manner and act in a manner consistent with consistent with human behavior form human behavior form the field known as the field known as artificial intelligence artificial intelligence . . The Turing Test The Turing Test A human questioner inputs A human questioner inputs questions and guesses which questions and guesses which respondent is human, based on respondent is human, based on the answers given. the answers given. If the questioner cannot tell If the questioner cannot tell which respondent is human, which respondent is human, the software passes the test. the software passes the test.

Upload: charlotte-dulcie-walton

Post on 27-Dec-2015

223 views

Category:

Documents


4 download

TRANSCRIPT

Page 1: Chapter 13: Artificial Intelligence Chapter 13 Artificial Intelligence Page 139 Computations that make it possible for a machine to perceive, reason,

Chapter 13: Artificial Chapter 13: Artificial IntelligenceIntelligence

Chapter 13Chapter 13Artificial Artificial

IntelligenceIntelligencePage Page 11

Computations that Computations that make it possible for a make it possible for a machine to perceive, machine to perceive, reason, and act in a reason, and act in a manner consistent manner consistent with human behavior with human behavior form the field known form the field known as as artificial artificial intelligenceintelligence..The Turing TestThe Turing TestA human questioner inputs A human questioner inputs

questions and guesses which questions and guesses which respondent is human, based respondent is human, based

on the answers given.on the answers given.If the questioner cannot tell If the questioner cannot tell which respondent is human, which respondent is human, the software passes the test.the software passes the test.

Page 2: Chapter 13: Artificial Intelligence Chapter 13 Artificial Intelligence Page 139 Computations that make it possible for a machine to perceive, reason,

Computer ReasoningComputer Reasoning

Chapter 13Chapter 13Artificial Artificial

IntelligenceIntelligencePage Page 22

To simulate logical reasoning, To simulate logical reasoning, heuristic functionsheuristic functions are are often used.often used.A heuristic is an artificial measure of how close the A heuristic is an artificial measure of how close the computer’s current status is to its problem-solving computer’s current status is to its problem-solving goal.goal.

In the tic-tac-toe example above, when the computer In the tic-tac-toe example above, when the computer is ready to make an X-move, it uses the heuristic is ready to make an X-move, it uses the heuristic max{max{FF((configconfig), where ), where configconfig can be the result of any can be the result of any O-move}.O-move}.

FF((configconfig) =) =

(# of completable rows, columns, and (# of completable rows, columns, and diagonals for X-player) – (# of diagonals for X-player) – (# of completable rows, columns, and completable rows, columns, and diagonals for O-player)diagonals for O-player)

if if configconfig is a non-winning is a non-winning configurationconfiguration

if if configconfig is an X-win is an X-win

-- if if configconfig is an O-win is an O-winO O

X

X

X O O

X

X

X O O

O X

X

X O O

XX O

X O O

X

X O

X O O

X O

X

2-1=1 3-1=2 2-1=1 3-1=2

O O

X X

X

O O O

X X

X

O O

X XX O

O O

X X

X O

O O

X X O

X

- 3-2=1 2-2=0 3-2=1

O O

X

X X

O O O

X

X X

O O

O XX X

O O

X

X X O

O O

X O

X X

- 2-2=0 2-2=0 3-2=1

O O

X X

X

O O O

X X

X

O O

O X XX

O O

X X

X O

O O

X X

X O

- 2-1=1 2-1=1 2-1=1

O O

X

X X

O O O

X

X X

O O

O XX X

O O

X

X O X

O O

X O

X X

- 2-1=1 3-1=2 3-1=2

1

- - -

-

Page 3: Chapter 13: Artificial Intelligence Chapter 13 Artificial Intelligence Page 139 Computations that make it possible for a machine to perceive, reason,

Expert SystemsExpert Systems

Chapter 13Chapter 13Artificial Artificial

IntelligenceIntelligencePage Page 33

By programming a computer with the By programming a computer with the assistance of experts in a particular field, assistance of experts in a particular field, an an expert systemexpert system can be developed to can be developed to perform very specialized tasks.perform very specialized tasks.

Users in need Users in need of expert of expert assistance assistance complete on-complete on-line line questionnaires questionnaires and the expert and the expert system system analyzes their analyzes their responses and responses and assigns assigns probabilities to probabilities to various various diagnoses.diagnoses.

Page 4: Chapter 13: Artificial Intelligence Chapter 13 Artificial Intelligence Page 139 Computations that make it possible for a machine to perceive, reason,

Neural NetworksNeural Networks

Chapter 13Chapter 13Artificial Artificial

IntelligenceIntelligencePage Page 44

To simulate learning, certain multiprocessor To simulate learning, certain multiprocessor systems, called systems, called neural networksneural networks, have been built to , have been built to “learn” to recognize particular patterns as correct “learn” to recognize particular patterns as correct or incorrect, based upon a trial-and-error process.or incorrect, based upon a trial-and-error process.In the example below, a neural network is used to In the example below, a neural network is used to teach a computerized system how to back a truck teach a computerized system how to back a truck up to a loading dock.up to a loading dock.

The physical The physical characteristics of the truck characteristics of the truck are programmed, with the are programmed, with the relationship between the relationship between the steering wheel, the tires, steering wheel, the tires, the cab, and the trailer the cab, and the trailer

formally calculated.formally calculated.

Starting at some initial Starting at some initial position, the truck is backed position, the truck is backed up one meter at a time, with up one meter at a time, with programmed steering; the programmed steering; the

error in the result is measured error in the result is measured and factored into the next and factored into the next attempt, until the error is attempt, until the error is

zero.zero.

Page 5: Chapter 13: Artificial Intelligence Chapter 13 Artificial Intelligence Page 139 Computations that make it possible for a machine to perceive, reason,

Natural Language ProcessingNatural Language Processing

Chapter 13Chapter 13Artificial Artificial

IntelligenceIntelligencePage Page 55

Written Comprehension: How can a Written Comprehension: How can a computer be programmed to grasp the computer be programmed to grasp the syntax and semantics of a natural syntax and semantics of a natural language?language?John saw the boy in the John saw the boy in the park with the telescope.park with the telescope.Question: Whose telescope is Question: Whose telescope is it?it?

Answer: John’sAnswer: John’s

John saw the boy in the John saw the boy in the park with the puppy.park with the puppy.

Question: Whose puppy is it?Question: Whose puppy is it?Answer: The boy’sAnswer: The boy’s

John saw the boy in the John saw the boy in the park with the statue.park with the statue.

Question: Whose statue is it?Question: Whose statue is it?Answer: The Answer: The park’spark’s

Page 6: Chapter 13: Artificial Intelligence Chapter 13 Artificial Intelligence Page 139 Computations that make it possible for a machine to perceive, reason,

Speech RecognitionSpeech Recognition

Chapter 13Chapter 13Artificial Artificial

IntelligenceIntelligencePage Page 66

00110101010000011010101000010111110101001011111010100100110101010010011010101001010111100100101011110010101101010011010110101001101010101010101101010101010101010101010010101010101001

1.1. The PC sound card The PC sound card converts analog sound converts analog sound waves spoken into a waves spoken into a microphone into a microphone into a digital format.digital format.

2.2. A A software software acoustical acoustical model model breaks breaks the word the word into into phonemesphonemes..

0011010101000 0011010101000 “K”“K”0101111101010 0101111101010 “AH”“AH”0100110101010 0100110101010 “M”“M”0101011110010 0101011110010 “P”“P”1011010100110 1011010100110 “Y”“Y”1010101010101 1010101010101 “OO”“OO”0101010101001 0101010101001 “T”“T”

3.3. A software A software language language model model compares compares the the phonemes phonemes to words in to words in its its dictionary.dictionary.

““K”K”““AHAH””““M”M”““P”P”““Y”Y”““OOOO””““T”T”

CACALMLMCOMCOMMAMACOMPCOMPAREARECOMPCOMPEETETECOMPCOMPLELETETECOMPUTECOMPUTE

4.4. Once the software decides on Once the software decides on the most likely candidate, it the most likely candidate, it displays that word.displays that word.

COMPUTCOMPUTEE

Page 7: Chapter 13: Artificial Intelligence Chapter 13 Artificial Intelligence Page 139 Computations that make it possible for a machine to perceive, reason,

RoboticsRobotics

Chapter 13Chapter 13Artificial Artificial

IntelligenceIntelligencePage Page 77

Robots are programmable devices capable of Robots are programmable devices capable of manipulating objects and performing tasks much manipulating objects and performing tasks much like humans are able to do.like humans are able to do.One of the more difficult problems when programming a One of the more difficult problems when programming a robot is determining when it is about to collide with robot is determining when it is about to collide with something, when it has collided with something, and what to something, when it has collided with something, and what to do in response to a collision.do in response to a collision.

Collision AvoidanceCollision Avoidance

Use scanners to Use scanners to determine the robot’s determine the robot’s

proximity to other proximity to other objects, redirecting the objects, redirecting the robot when a collision robot when a collision

is imminent.is imminent.

Collision DetectionCollision Detection

Use sensors at Use sensors at strategically strategically

located places on located places on the robot to the robot to

determine if a determine if a collision occurs.collision occurs.

Collision Collision ReactionReaction

Go around?Go around?Climb over?Climb over?

Bounce back?Bounce back?Run away?Run away?Drop dead?Drop dead?

Page 8: Chapter 13: Artificial Intelligence Chapter 13 Artificial Intelligence Page 139 Computations that make it possible for a machine to perceive, reason,

Robot ChallengesRobot Challenges

Chapter 13Chapter 13Artificial Artificial

IntelligenceIntelligencePage Page 88

Other common human actions that are difficult to Other common human actions that are difficult to program include propelled locomotion and manual program include propelled locomotion and manual manipulation.manipulation.

Walking GaitWalking Gait

How can a robot be How can a robot be programmed to propel programmed to propel itself forward on “legs” itself forward on “legs”

and still maintain its and still maintain its balance?balance?

GraspingGrasping

How can a robot be How can a robot be programmed to grasp part programmed to grasp part

of a stack of objects, of a stack of objects, without toppling the rest of without toppling the rest of

the stack?the stack?

Page 9: Chapter 13: Artificial Intelligence Chapter 13 Artificial Intelligence Page 139 Computations that make it possible for a machine to perceive, reason,

AI in GamesAI in Games

Chapter 13Chapter 13Artificial Artificial

IntelligenceIntelligencePage Page 99

Non-player characters in games need to appear Non-player characters in games need to appear intelligent, even though they are controlled by the intelligent, even though they are controlled by the game program instead of by a game player.game program instead of by a game player.

FlockingFlockingBy providing a group of characters By providing a group of characters with simple goals and behaviors, a with simple goals and behaviors, a

“mob mentality” can be “mob mentality” can be implemented with a minimum of implemented with a minimum of

code.code.

Dead ReckoningDead ReckoningBy having a game-By having a game-

driven predator driven predator character react to the character react to the anticipated position of anticipated position of its player-driven prey its player-driven prey

(using the prey’s (using the prey’s current position and current position and

velocity), a chase can velocity), a chase can appear more realistic.appear more realistic.