Transcript Document

ARTIFICIAL INTELLIGENCE
IS 340
CHANDRA S. AMARAVADI
1 of 45
ARTIFICIAL INTELLIGENCE
IN THIS PRESENTATION
Introduction to AI
Milestones & early work
Machine Intelligence
The Nature of knowledge
Knowledge representation
Examples
Neural nets
Business & recent applications
2 of 45
INTRODUCTION TO AI
3 of 45
THE HISTORY OF AI (FYI)
Major milestones
•Alan Turing & test for intelligence
•AI as a field of study
•Lisp language
•Expert Systems
•Dendral & Mycin
•Small Talk, Prolog
•Fifth Generation Project
•Honda robot
•Stanford driverless car
-----
1950
1956
1958
1965
-----
1972
1981
1995
2005
4 of 45
EARLY RESEARCH
Early research on AI focussed on:
Logic
Perceptrons
Chess
Blocks world (a world consisting of only blocks)
5 of 45
SEARCH STRATEGIES
Generate and Test
Generate a possible solution
and test to see if it is the answer


Breadth-first
Depth-first
 Heuristic
 Hill-climbing
?
?
?
6 of 45
DEFINING INTELLIGENCE
7 of 45
DEFINITION
Artificial Intelligence (AI)
AI is concerned with the principles and mechanisms
for achieving intelligent behavior in machines
8 of 45
BRANCHES OF AI
Artificial
intelligence
Expert
Systems
NLP
Robotics
Vision
Systems
Machine
Learning
9 of 45
NATURE OF INTELLIGENCE
Knowledge + Reasoning power
= Intelligence
Any other method of achieving intelligence?
10 of 45
Top-down - build logical equivalents, e.g. LOGIC,
Expert systems
Bottom-up - build physical equivalents, e.g.
perceptrons, neural nets
11 of 45
THE TEST FOR MACHINE INTELLIGENCE
The Turing test: If a person interacting with an entity from a remote
location is unable to judge whether he/she is dealing with a computer or
a human, and the entity a machine, it is said to possess intelligence.
Questions
?
Responses
12 of 45
THE NATURE OF
KNOWLEDGE
13 of 45
KNOWLEDGE
Knowledge: information organized for
problem solving
facts,
constraints,
problems,
goals,
procedures.
14 of 45
THE NATURE OF KNOWLEDGE
Two types of knowledge:
Declarative – Knowledge about an object (size, shape etc.)
Procedural – Knowledge about how to do something.
(how to install memory)
15 of 45
KNOWLEDGE REPRESENTATION
A Sampling of Knowledge
How to install a water pump
The definition of a “field goal”
Painters & styles from the modern era
The process of becoming a GSA contractor
The architectural differences between AMD &
Intel chips
The meaning of “Lousiana report” in the context
of a faculty committee meeting.
16 of 45
KNOWLEDGE
REPRESENTATION
17 of 45
KNOWLEDGE REPRESENTATION
Knowledge representation is concerned with
how to encode knowledge
Logic (Predicate logic)
Frames
Scripts
Semantic nets (Snets)
Rules
18 of 45
IDENTIFY THESE AS EXAMPLES
OF LOGIC, FRAMES, SCRIPTS…
EXAMPLE 1
sister_of(X,Y), bird_of_prey(X),
father_of(robin, Y)
father_of(robin,_)
EXAMPLE 2
EXAMPLE 3
If # of users > 300 then,
license fee = $500
If # of users < 300 then,
license fee = $300
is_a
: dbms
software cost
: $3,000
License cost
: check_with_vendor
no of users
: 2000
Max # of tables : 10,000
Supports ODBC : Yes
19 of 45
EXAMPLES OF KNOWLEDGE
REPRESENTATIONS..
EXAMPLE 5
EXAMPLE 4
Bird-of-prey
P PTRANS P to P.O.
P ATTEND eyes to counter
Is-a
P MBUILD line position
P PTRANS P to line
Eagle
P PTRANS M to X
X PTRANS Stamps to P
Max
Wingspan
1.5 m
Bird
Is-a
Max
Speed
20 Knots
20 of 45
NOTES ON SEMANTIC NETS
 Based on associative memory
 “node” + “link” formalism
 nodes represent concepts or values
 links can be structural or descriptive
 represent structure or characteristic
21 of 45
NOTES ON RULES
 Origins in S-R paradigms
 Thought to be used by experts
 Have a IF…THEN… format
Note: S-R: stimulus/response
22 of 45
NOTES ON SCRIPTS
 A description (conceptual representation) of
actions in a pre-defined situation
 Originated from film industry
 Consists of actors/props
 Act in predictable ways
23 of 45
EXAMPLE OF LOGIC
facts:
has_qualification(brad,3.2,620).
has_qualification(jill,4.0,540).
has_qualification(ted,3.5,320).
has_qualification(matt,3.8, 600).
Predicates:
select(X) :- has_qualification(X,GPA,GMAT),
GPA>3.2, GMAT>550;
Goals:
select(brad)? jill? ted? matt?
24 of 45
FOR DISCUSSION
Identify whether the following types of knowledge are declarative or
procedural and identify a suitable representation scheme, give rationale:
1. Admit students to MBA program if they have a gmat score of
> 550
2. A description of computing facilities at WIU.
3. A proof of the theorem that any triangle circumscribed by a
semi-circle will always be a right angled triangle
4. Instructions for assembling a PC
5. Family relationships -- X and Y are the parents of P & Q; P has a
maternal aunt Z.
6. Stages in a software life cycle -- analysis, design, implementation etc.
25 of 45
NEURAL NETS
Mathematical models to simulate neural models of the brain,
Often used in applications requiring pattern recognition e.g.
crime, fraud, intrusion detection etc.
Neurons
Dendrites
The brain
nose
hair color
eyes
gait
Neural Net
(a math model)
26 of 45
BUSINESS APPLICATIONS OF AI
Automated voice response
Text mining
Production applications
machine design
robotics
paper thickness
Scheduling of cranes
Credit approval
27 of 45
INDUSTRIAL APPLICATIONS OF AI
Driverless vehicles
Facial recognition
Crime prevention
Pothole recognition
Drones
28 of 45
Can a machine ever have the intelligence of a human
being?
 Has Turing’s test been passed?
 Why did early researchers concentrate on Chess?
 If we make use of a frog’s brain to process stimuli, is that
an example of a Top-Down or a Bottom-up approach?
 What branch of AI does the work on perceptrons
resemble?
 What “hardware” item is essential equipment for vision
systems?
 Are robots useful in industry? How?
 If a machine is taking dictation, is it necessary to
understand the text or can it be done mechanically?

29 of 45
The End!
Please note there are only 29 slides
30 of 45