2. Goals of this Course
This class is a broad introduction to artificial
intelligence (AI)
AI is a very broad field with many subareas
We will cover many of the primary concepts/ideas
But in 10 weeks we can’t cover everything
3. Why taking ICS 3211could change
your life…..
As we begin the new millennium
science and technology are changing rapidly
“old” sciences such as physics are relatively well-understood
computers are ubiquitous
Grand Challenges in Science and Technology
understanding the brain
reasoning, cognition, creativity
creating intelligent machines
is this possible?
what are the technical and philosophical challenges?
arguably AI poses the most interesting challenges and
questions in computer science today
4. Today’s Lecture
What is intelligence? What is artificial intelligence?
A very brief history of AI
Modern successes: Stanley the driving robot; Deep Blue the
chess game player; IBM’s Watson computer
An AI scorecard
How much progress has been made in different aspects of AI
AI in practice
Successful applications
The rational agent view of AI
5. What is Intelligence?
Intelligence:
“the capacity to learn and solve problems” (Websters
dictionary)
in particular,
the ability to solve novel problems
the ability to act rationally
the ability to act like humans
Artificial Intelligence
build and understand intelligent entities or agents
2 main approaches: “engineering” versus “cognitive
modeling”
6. Types of Intelligence
According to Howard Gardner’s multiple intelligence theory,
there are various types of intelligence viz:
General intelligence: -
Abilities that allow us to be flexible and adaptive thinkers, not
necessarily tied to acquired knowledge.
Linguistic-verbal intelligence: -
Use words and language in various forms / Ability to manipulate
language to express oneself poetically
Logical-Mathematical intelligence: -
Ability to detect patterns / Approach problems logically / Reason
deductively
Musical intelligence: -
Recognize nonverbal sounds: pitch, rhythm, and tonal patterns
Spatial intelligence: -
Typically thinks in images and pictures / Used in both arts and
sciences
7. Types of Intelligence (2)
Intrapersonal intelligence: -
Ability to understand oneself, including feelings and
motivations / Can discipline themselves to accomplish a
wide variety of tasks
Interpersonal intelligence: -
Ability to "read people"—discriminate among other
individuals especially their moods, intentions, motivations;
/ Adept at group work, typically assume a leadership role.
Naturalist intelligence: -
Ability to recognize and classify living things like plants,
animals
Bodily-Kinesthetic intelligence: -
Use one’s mental abilities to coordinate one’s own bodily
movements
8. Types of Intelligence (3)
Note:
Understanding the various types of intelligence
provides theoretical foundations for recognizing
different talents and abilities in people
"What makes life interesting, however, is that we
don’t have the same strength in each intelligence
area, and we don’t have the same amalgam of
intelligences. Just as we look different from one
another and have different kinds of personalities,
we also have different kinds of minds."
9. Signs of Intelligence
Learn or understand from experience
Make sense out of ambiguous or contradictory
messages
Respond quickly and successfully to new
situations
Use reasoning to solve problems
10. More Signs of Intelligence
Deal with perplexing situations
Understand and infer in ordinary, rational ways
Apply knowledge to manipulate the
environment
Think and reason
Recognize the relative importance of different
elements in a situation
11. What is AI? (2)
There is no agreed definition of the term artificial intelligence.
However, there are various definitions that have been
proposed. These are considered below.
AI is a study in which computer systems are made that think like
human beings. Haugeland, 1985 & Bellman, 1978.
AI is a study in which computer systems are made that act like
people.
AI is the art of creating computers that perform functions that
require intelligence when performed by people. Kurzweil, 1990.
AI is the study of how to make computers do things, which at the
moment people are better at. Rich & Knight
AI is the study of computations that make it possible to perceive,
reason and act. Winston, 1992
AI is considered to be a study that seeks to explain and emulate
intelligent behaviour in terms of computational processes.
Schalkeoff, 1990.
AI is considered to be a branch of computer science that is
concerned with the automation of intelligent behavior. Luger &
Stubblefield, 1993.
12. What is AI? (3)
Artificial Intelligence is the development of
systems that exhibit the characteristics we
associate with intelligence in human behavior:
perception,
natural language processing,
reasoning,
planning,
problem solving,
learning and adaptation,
etc.
13. What’s involved in Intelligence?
Ability to interact with the real world
to perceive, understand, and act
e.g., speech recognition and understanding and synthesis
e.g., image understanding
e.g., ability to take actions, have an effect
Reasoning and Planning
modeling the external world, given input
solving new problems, planning, and making decisions
ability to deal with unexpected problems, uncertainties
Learning and Adaptation
we are continuously learning and adapting
our internal models are always being “updated”
e.g., a baby learning to categorize and recognize animals
14. Academic Disciplines relevant to
AI
Philosophy Logic, methods of reasoning, mind as physical
system, foundations of learning, language,
rationality.
Mathematics Formal representation and proof, algorithms,
computation, (un)decidability, (in)tractability
Probability/Statistics modeling uncertainty, learning from data
Economics utility, decision theory, rational economic agents
Neuroscience neurons as information processing units.
Psychology/ how do people behave, perceive, process cognitive
Cognitive Science information, represent knowledge.
Computer building fast computers
engineering
Control theory design systems that maximize an objective
function over time
Linguistics knowledge representation, grammars
15. AI Problems
What are problems contained in AI?
Much of the early work focused on formal tasks,
such as game playing and theorem proving.
E.g. Chess playing, Logic Theorist was an early
attempt to prove mathematical theorems.
Game playing and theorem proving share the
property that people who do them well are
considered to be displaying intelligence.
16. AI Problems-Many solution
alternatives
Despite this it appeared that computers could
perform well at those tasks by being fast at
exploring a large number of solution paths and
then selecting the best one.
No computer is fast enough to overcome the
combinatorial explosion generated by most
problems.
17. AI Problems- Based on
commonsense reasoning
AI focusing on the sort of problem solving we
do every day for instance when we decide to
get to work in the morning, often called
commonsense reasoning.
In investigating this sort of reasoning Newell,
Shaw, and Simon built the General Problem
Solver (GPS), which they applied to several
commonsense tasks as well performing symbolic
manipulations of logical expression
18. AI Problems-with domain
knowledge
However no attempt was made to create a
program with a large amount of knowledge
about a particular problem domain. Only quite
simple tasks were selected.
As AI research progressed techniques for
handling larger amounts of world knowledge
were developed to deal with problem solving in
specialised domains such as medical
diagnosis and chemical analysis.
19. AI Problems- perception via
speech, vision, and natural
language
Perception (vision and speech) is another area
for AI problems.
Natural language understanding, and problem
solving in specialised domain are other areas
related to AI problems.
20. AI Problems- Speech based
The problem of understanding spoken
language is perceptual problem and is hard to
solve from the fact that it is more analog
related than digital related.
Many people can perform one or may be more
specialised tasks in which carefully acquired
expertise is necessary.
21. AI Problems-Tasks
Examples of such tasks include engineering
design, scientific discovery, medical diagnosis,
and financial planning.
Programs that can solve problems in these
domains also fall under the aegis of Artificial
Intelligence.
22. AI Problems- List of tasks
List of the tasks that are targets of works in AI
are:-
Mundane tasks
Perception
Vision
Speech
Natural Language
Understanding
Generation
Translation
Commonsense reasoning
Robot Control
23. AI Problems- List of Tasks (2)
Targets of AI (cont….):
Formal Tasks
Games
Chess, etc.
Mathematics
Geometry, Logic, Integral Calculus, etc.
Expert Tasks
Engineering
Design, Fault Finding, Manufacturing Planning
Scientific analysis
Medical Diagnosis
Financial Analysis
24. AI Problems- Problem-solving
Strategy
A person who knows how to perform tasks
from several of the categories shown in above
list learn the necessary skills in a standard
order.
First perceptual, linguistic, and commonsense
skills are learned.
Later expert skills such as engineering, medicine,
or finance are acquired
25. AI Problems-Problem-solving
Strategy
Earlier skills are easier and thus more
amenable to computerised duplication than the
later, more specialised one.
For this reason much of the initial work in AI
work was concentrated in those early areas.
26. AI Problems- Success
Scenarios
The problems areas where now AI is
flourishing most as a practical discipline are
primarily the domains that require only
specialised expertise without the assistance of
commonsense knowledge.
Expert systems (AI programs) now are up for
day-to-day tasks that aim at solving part, or
perhaps all, of practical, significant problem
that previously required high human expertise.
27. AI Problems- Fundamental
Questions
The following four questions need to be
considered before we can progress further:
1) What are the underlying assumptions about
intelligence?
2) What kinds of techniques will be useful for
solving AI problems?
3) At what level if at all can human intelligence be
modelled?
4) When will it be realised when an intelligent
program has been built?
28. QUESTION#1: The Underlying
Assumption
A physical symbol system consists of a set of
entities called symbols which are patterns that
can occur as components of another entitiy
called an expression.
At an instant the system will contain a collection
of these symbol structures. In addition the
system also contains a collection of processes
that operate on expressions to produce other
expressions; processes of creation,
modification, reproduction and destruction.
29. QUESTION#1: The Underlying
Assumption
A physical symbol system is a machine that
produces through time an evolving collection
of symbol structures. Such a system is a
machine that produces through time an
evolving collection of symbol structures.
AI system is a physical symbol
system.
30. Example of Physical Symbol
Systems
Formal logic: the symbols are words like
"and", "or", "not", "for all x" and so on. The
expressions are statements in formal logic
which can be true or false. The processes
are the rules of logical deduction.
Algebra: the symbols are "+", "×", "x", "y",
"1", "2", "3", etc. The expressions are
equations. The processes are the rules of
algebra, that allow you to manipulate a
mathematical expression and retain its truth.
31. Example of Physical Symbol
Systems
A digital computer: the symbols are zeros and
ones of computer memory, the processes are
the operations of the CPU that change
memory.
Chess: the symbols are the pieces, the
processes are the legal chess moves, the
expressions are the positions of all the pieces
on the board
32. Physical Symbol Systems
The physical symbol system hypothesis
claims that both of these are also examples of
physical symbol systems
Intelligent human thought: the symbols are
encoded in our brains. The expressions are
thoughts. The processes are the mental
operations of thinking.
A running artificial intelligence program: The
symbols are data. The expressions are more
data. The processes are programs that
manipulate the data.
33. The Importance of hypothesis
The importance of the physical symbol system
hypothesis is twofold.
It is significant theory about the nature of human
intelligence.
Forms the basis of the belief that it is possible to
build programs that can perform intelligent tasks
now performed by people.
34. QUESTION#2: An AI Technique
Intelligence requires knowledge but knowledge
possesses less desirable properties such as
a) it is voluminous
b) it is difficult to characterize accurately
c) it is constantly changing
d) it differs from data by being organized in a way
that corresponds to its application
35. QUESTION#2: AI Technique(2)
An AI technique is a method that exploits
knowledge that is represented so that
The knowledge captures generalisations ;
situations that share properties, are grouped
together, rather than being allowed separate
representation
36. QUESTION#2: AI Technique(3)
It can be understood by people who must provide
it; although for many programs the bulk of the
data may come automatically, such as from
readings.
In many AI domains people must supply the
knowledge to programs in a form the people
understand and in a form that is acceptable to the
program.
It can be easily modified to correct errors and
reflect changes in real conditions
37. QUESTION#2: AI Technique(4)
It can be widely used even if it is incomplete or
inaccurate.
It can be used to help overcome its own sheer
bulk by helping to narrow the range of possibilities
that must be usually considered.
In order to characterize an AI technique let us
consider a question answering example.
38. Example: Question Answering
Lets discuss a program that reads in English
Text and answers questions, also stated in
English.
This kind of task is difficult to state formally
and precisely what our problem is and what
constitutes correct solutions to it.
39. Example: Question
Answering(2)
For example suppose that the input text were
just the single sentence:
Russia massed troops on the Ukrainian
border
Then either of the following question-
answering dialogues might occur in the
POLITICS program.
40. Example: Question
Answering(3)
Dialogue 1
Q: Why did Russia do this?
A: Because Russia thought that it could take
political control of Ukraine by sending troops.
Q: What should the United States do?
A: The United States should intervene
militarily.
41. Example: Question
Answering(4)
Dialogue 2
Q: Why did Russia do this?
A: Because Russia wanted to increase its
political influence over Ukraine.
Q: What should the United States do?
A: The United States should denounce the
Russian action in the United Nations
42. Example: Question
Answering(5)
In the POLITICS program answers were
constructed by considering both the input text
and a separate model of the beliefs and
actions of various political entities including
Russia.
When the model is changed, as it was
between these two dialogues, the system’s
answers also change.
43. Example: Question
Answering(6)
In the above example, the first dialogue was
produced when POLITICS was given a model
of conservatives (e.g. views of the Republican
party).
The second dialogue occurred when
POLITICS program was given a model that
was intended to correspond to the beliefs of a
typical American liberal(e.g. Democratic party).
44. Example: Question
Answering(7)
The general point is that defining what it
means to produce a correct answer to a
question may be very hard.
Usually question-answering form defines what
it means by to be an answer by the procedure
that is used to compute the answer.
Then their authors appeal to other people to
agree that the answers found by the program
“make sense”.
45. Conclusion: AI Technique to
Apply
Three important AI techniques are
Search – Provides a way of solving problems for
which no more direct approach is available.
Various solution alternatives have to be explored e.g.
state-space search, adversarial search(aka game
playing)
Use of Knowledge – Provides a way of solving
complex problems by exploiting the structures of
the objects that are involved.
Abstraction – Provides a way of separating
important features and variations from the many
unimportant ones that would otherwise
overwhelm any process.
46. QUESTION#3: The Level of the
Model
Before starting doing something, it is good
idea to decide exactly what we are trying to do.
We must ask ourselves:
What is our goal in trying to produce programs
that do the tasks the same way people do?
Are we trying to produce programs that do the
tasks the same way people do?
Or are we trying to produce programs that simply
do the tasks in whatever way appears easiest?
47. QUESTION#3: The Level of the
Model(2)
Efforts to build program that perform tasks the
way people do can be divided into two classes
Those that attempt to solve problems that do
not really fit our definition of AI. i.e. problems
that computer could easily solve.
48. QUESTION#3: The Level of the
Model(3)
The second class attempt to model human
performance are those that do things that fall
more clearly within our definition of AI tasks;
they do things that are not trivial for the computer.
Reasons for modeling human performance at
these kind of tasks:-
49. QUESTION#3: The Level of the
Model(4)
To test psychological theories of human
performance. E.g. PARRY program written for this
reason, which exploited a model of human paranoid
behaviour to simulate the conversational behaviour
of a paranoid person.
To enable computer to understand human
reasoning. For example, for a computer to be able
to read a news paper story and then answer
question, such as “Why did Brazil lose the game?”
50. QUESTION#3: The Level of the
Model(5)
To enable people to understand computer
reasoning. In many cases people are reluctant
to rely on the output of computer unless they
can understand how the machine arrived at its
result.
To exploit what knowledge we can collect from
people.
To ask for assistance from best performing people
and ask them how to proceed in dealing with their
tasks.
51. QUESTION#4: Criteria for
success.
One of the most important questions to answer
in any scientific or engineering research
project is “How will we know if we have
succeeded?”
So in AI we have to ask ourselves how will we
know if we have constructed a machine that is
intelligent?
The question is as hard as unanswerable
question “What is Intelligence?”
52. QUESTION#4: Criteria for
success(2)
To measure the progress we use proposed
method known as Turing Test.
Alan Turing suggested this method to
determine whether the machine can think.
To conduct this test, we need two people and the
machine to be evaluated. One person act as
interrogator, who is in a separate room from the
computer and the other person.
53. QUESTION#4: Criteria for
success(3)
The interrogator can ask questions of either the
person or computer by typing questions and
receiving typed responses
However the interrogator knows them only as A
and B and aims to determine which is the person
and which is the machine.
The goal of the machine is to fool the interrogator
into believing that it is the person. If the machine
succeeds at this, then we will conclude that the
machine can think.
54. History of AI
1943: early beginnings
McCulloch & Pitts: Boolean circuit model of brain
1950: Turing
Turing's "Computing Machinery and Intelligence“
1956: birth of AI
Dartmouth meeting: "Artificial Intelligence“ name adopted
1950s: initial promise
Early AI programs, including
Samuel's checkers program
Newell & Simon's Logic Theorist
1955-65: “great enthusiasm”
Newell and Simon: GPS, general problem solver
Gelertner: Geometry Theorem Prover
McCarthy: invention of LISP
55. History of AI
1966—73: Reality dawns
Realization that many AI problems are intractable
Limitations of existing neural network methods identified
Neural network research almost disappears
1969—85: Adding domain knowledge
Development of knowledge-based systems
Success of rule-based expert systems,
E.g., DENDRAL, MYCIN
But were brittle and did not scale well in practice
1986-- Rise of machine learning
Neural networks return to popularity
Major advances in machine learning algorithms and applications
1990-- Role of uncertainty
Bayesian networks as a knowledge representation framework
1995-- AI as Science
Integration of learning, reasoning, knowledge representation
AI methods used in vision, language, data mining, etc
56. Success Stories
Deep Blue defeated the reigning world chess champion Garry
Kasparov in 1997
AI program proved a mathematical conjecture (Robbins conjecture)
unsolved for decades
During the 1991 Gulf War, US forces deployed an AI logistics
planning and scheduling program that involved up to 50,000
vehicles, cargo, and people
NASA's on-board autonomous planning program controlled the
scheduling of operations for a spacecraft
Proverb solves crossword puzzles better than most humans
Robot driving: DARPA grand challenge 2003-2007
2006: face recognition software available in consumer cameras
57. Example: DARPA Grand
Challenge
Grand Challenge
Cash prizes ($1 to $2 million) offered to first robots to complete a long course
completely unassisted
Stimulates research in vision, robotics, planning, machine learning, reasoning,
etc
2004 Grand Challenge:
150 mile route in Nevada desert
Furthest any robot went was about 7 miles
… but hardest terrain was at the beginning of the course
2005 Grand Challenge:
132 mile race
Narrow tunnels, winding mountain passes, etc
Stanford 1st, CMU 2nd, both finished in about 6 hours
2007 Urban Grand Challenge
November in Victorville, California
58. HAL: from the movie 2001
2001: A Space Odyssey
classic science fiction movie from 1969
HAL
part of the story centers around an intelligent computer called
HAL
HAL is the “brains” of an intelligent spaceship
in the movie, HAL can
speak easily with the crew
see and understand the emotions of the crew
navigate the ship automatically
diagnose on-board problems
make life-and-death decisions
display emotions
In 1969 this was science fiction: is it still science fiction?
59. Hal and AI
HAL’s Legacy: 2001’s Computer as Dream and Reality
MIT Press, 1997, David Stork (ed.)
discusses
HAL as an intelligent computer
are the predictions for HAL realizable with AI today?
Materials online at
http://mitpress.mit.edu/e-books/Hal/contents.html
The website contains
full text and abstracts of chapters from the book
links to related material and AI information
sound and images from the film
60. Consider what might be involved
in building a computer like Hal….
What are the components that might be
useful?
Fast hardware?
Chess-playing at grandmaster level?
Speech interaction?
speech synthesis
speech recognition
speech understanding
Image recognition and understanding ?
Learning?
Planning and decision-making?
62. Can we build hardware as complex as the
brain?
How complicated is our brain?
a neuron, or nerve cell, is the basic information processing unit
estimated to be on the order of 10 12 neurons in a human brain
many more synapses (10 14) connecting these neurons
cycle time: 10 -3 seconds (1 millisecond)
How complex can we make computers?
108 or more transistors per CPU
supercomputer: hundreds of CPUs, 1012 bits of RAM
cycle times: order of 10 - 9 seconds
Conclusion
YES: in the near future we can have computers with as many basic
processing elements as our brain, but with
far fewer interconnections (wires or synapses) than the brain
much faster updates than the brain
but building hardware is very different from making a computer behave like a
brain!
63. Can Computers Talk?
This is known as “speech synthesis”
translate text to phonetic form
e.g., “fictitious” -> fik-tish-es
use pronunciation rules to map phonemes to actual sound
e.g., “tish” -> sequence of basic audio sounds
Difficulties
sounds made by this “lookup” approach sound unnatural
sounds are not independent
e.g., “act” and “action”
modern systems (e.g., at AT&T) can handle this pretty well
a harder problem is emphasis, emotion, etc
humans understand what they are saying
machines don’t: so they sound unnatural
Conclusion:
NO, for complete sentences
64. Can Computers Recognize
Speech?
Speech Recognition:
mapping sounds from a microphone into a list of
words
classic problem in AI, very difficult
“Lets talk about how to wreck a nice beach”
(I really said “________________________”)
Recognizing single words from a small
vocabulary
systems can do this with high accuracy (order of 99%)
e.g., directory inquiries
limited vocabulary (area codes, city names)
computer tries to recognize you first, if unsuccessful hands
you over to a human operator
saves millions of dollars a year for the phone companies
65. Recognizing human speech
(ctd.)
Recognizing normal speech is much more difficult
speech is continuous: where are the boundaries between words?
e.g., “John’s car has a flat tire”
large vocabularies
can be many thousands of possible words
we can use context to help figure out what someone said
e.g., hypothesize and test
try telling a waiter in a restaurant:
“I would like some dream and sugar in my coffee”
background noise, other speakers, accents, colds, etc
on normal speech, modern systems are only about 60-70%
accurate
Conclusion:
NO, normal speech is too complex to accurately recognize
YES, for restricted problems (small vocabulary, single speaker)
66. Can Computers Understand
speech?
Understanding is different to recognition:
“Time flies like an arrow”
assume the computer can recognize all the words
how many different interpretations are there?
67. Can Computers Understand
speech?
Understanding is different to recognition:
“Time flies like an arrow”
assume the computer can recognize all the words
how many different interpretations are there?
1. time passes quickly like an arrow?
2. command: time the flies the way an arrow times the flies
3. command: only time those flies which are like an arrow
4. “time-flies” are fond of arrows
68. Can Computers Understand
speech?
Understanding is different to recognition:
“Time flies like an arrow”
assume the computer can recognize all the words
how many different interpretations are there?
1. time passes quickly like an arrow?
2. command: time the flies the way an arrow times the flies
3. command: only time those flies which are like an arrow
4. “time-flies” are fond of arrows
only 1. makes any sense,
but how could a computer figure this out?
clearly humans use a lot of implicit commonsense knowledge in
communication
Conclusion: NO, much of what we say is beyond the
capabilities of a computer to understand at present
69. Can Computers “see”?
Recognition v. Understanding (like Speech)
Recognition and Understanding of Objects in a scene
look around this room
you can effortlessly recognize objects
human brain can map 2d visual image to 3d “map”
Why is visual recognition a hard problem?
Conclusion:
mostly NO: computers can only “see” certain types of objects under limited
circumstances
YES for certain constrained problems (e.g., face recognition)
70. Can Computers Learn and
Adapt ?
Learning and Adaptation
consider a computer learning to drive on the freeway
we could teach it lots of rules about what to do
or we could let it drive and steer it back on course when it heads for
the embankment
systems like this are under development (e.g., Daimler Benz)
e.g., RALPH at CMU
in mid 90’s it drove 98% of the way from Pittsburgh to San Diego without any
human assistance
machine learning allows computers to learn to do things without
explicit programming
many successful applications:
requires some “set-up”: does not mean your PC can learn to forecast the
stock market or become a brain surgeon
Conclusion: YES, computers can learn and adapt, when
presented with information in the appropriate way
71. Can computers plan and make optimal
decisions?
Intelligence
involves solving problems and making decisions and plans
e.g., you want to take a holiday in Brazil
you need to decide on dates, flights
you need to get to the airport, etc
involves a sequence of decisions, plans, and actions
What makes planning hard?
the world is not predictable:
your flight is canceled or there’s a backup on the 405
there are a potentially huge number of details
do you consider all flights? all dates?
no: commonsense constrains your solutions
AI systems are only successful in constrained planning problems
Conclusion: NO, real-world planning and decision-making is still beyond the
capabilities of modern computers
exception: very well-defined, constrained problems
72. Summary of State of AI Systems in
Practice
Speech synthesis, recognition and understanding
very useful for limited vocabulary applications
unconstrained speech understanding is still too hard
Computer vision
works for constrained problems (hand-written zip-codes)
understanding real-world, natural scenes is still too hard
Learning
adaptive systems are used in many applications: have their limits
Planning and Reasoning
only works for constrained problems: e.g., chess
real-world is too complex for general systems
Overall:
many components of intelligent systems are “doable”
there are many interesting research problems remaining
73. Intelligent Systems in Your
Everyday Life
Post Office
automatic address recognition and sorting of mail
Banks
automatic check readers, signature verification systems
automated loan application classification
Customer Service
automatic voice recognition
The Web
Identifying your age, gender, location, from your Web surfing
Automated fraud detection
Digital Cameras
Automated face detection and focusing
Computer Games
Intelligent characters/agents
74. AI Applications: Machine
Translation
Language problems in international business
e.g., at a meeting of Japanese, Korean, Vietnamese and Swedish investors, no common
language
or: you are shipping your software manuals to 127 countries
solution; hire translators to translate
would be much cheaper if a machine could do this
How hard is automated translation
very difficult! e.g., English to Russian
“The spirit is willing but the flesh is weak” (English)
“the vodka is good but the meat is rotten” (Russian)
not only must the words be translated, but their meaning also!
is this problem “AI-complete”?
Nonetheless....
commercial systems can do a lot of the work very well (e.g.,restricted vocabularies in
software documentation)
algorithms which combine dictionaries, grammar models, etc.
Recent progress using “black-box” machine learning techniques
76. What’s involved in Intelligence?
(again)
Perceiving, recognizing, understanding the real
world
Reasoning and planning about the external world
Learning and adaptation
So what general principles should we use to
achieve these goals?
77. Characteristics of AI
Symbolic Processing
AI emphasizes manipulation of symbols rather than
numbers.
The manner in which symbols are processed is non-
algorithmic since most human reasoning process do
not necessarily follow a step by step approach
(algorithmic approach).
Heuristics
Are similar to rules of thumb where you need not
rethink completely what to do every time a similar
problem is encountered.
Inferencing
This is a form of reasoning with facts and rules using
heuristics or some search strategies.
78. Characteristics of AI (2)
Pattern matching
A process of describing objects, events or
processes in terms of their qualitative features and
logical and computational relationships.
Knowledge Processing
Knowledge consists of facts, concepts, theories,
heuristics methods, procedures and relationships.
Knowledge bases.
Collection of knowledge related to a problem or an
opportunity used in problem.
Reasoning occurs based on this knowledge base.
79. Contrasting AI with Natural
Intelligence
Important commercial advantages of AI are:-
1) AI is permanent as long as computer system and
programs remain unchanged
2) AI offers ease of duplications and dissemination as
compared to long apprenticeship for natural
intelligence.
3) AI can be less expensive than natural intelligence.
4) AI being a computer system is consistent and
thorough; natural intelligence may be erratic since
people are erratic, they don’t perform consistently.
5) AI can execute certain tasks much faster than
humans can.
6) AI can perform certain tasks better than many or
even most people.
80. Contrasting AI with Natural
Intelligence (2)
Natural Intelligence has the following
advantages
1) Natural intelligence is creative, AI is
uninspired- human always determine
knowledge.
2) Natural intelligence enables people to
benefit from use of sensory experience
directly, while most AI systems must work
with symbolic knowledge.
81. Different Types of Artificial
Intelligence
1. Modeling exactly how humans actually think
2. Modeling exactly how humans actually act
3. Modeling how ideal agents “should think”
4. Modeling how ideal agents “should act”
Modern AI focuses on the last definition
we will also focus on this “engineering” approach
success is judged by how well the agent performs
82. Thinking humanly
Cognitive Science approach
Try to get “inside” our minds
E.g., conduct experiments with people to try to “reverse-
engineer” how we reason, learning, remember, predict
Problems
Humans don’t behave rationally
e.g., insurance
The reverse engineering is very hard to do
The brain’s hardware is very different to a computer
program
83. Thinking rationally
Represent facts about the world via logic
Use logical inference as a basis for reasoning about these
facts
Can be a very useful approach to AI
E.g., theorem-provers
Limitations
Does not account for an agent’s uncertainty about the world
E.g., difficult to couple to vision or speech systems
Has no way to represent goals, costs, etc (important aspects of
real-world environments)
84. Acting rationally
Decision theory/Economics
Set of future states of the world
Set of possible actions an agent can take
Utility = gain to an agent for each action/state pair
An agent acts rationally if it selects the action that
maximizes its “utility”
Or expected utility if there is uncertainty
Emphasis is on autonomous agents that behave
rationally (make the best predictions, take the best
actions)
on average over time
within computational limitations (“bounded rationality”)
86. Symbolic AI
Based on Newell & Simons Physical Symbol System Hypothesis
Uses logical operations that are applied to declarative knowledge
bases (FOPL)
Commonly referred to as “Classical AI”
Represents knowledge about a problem as a set of declarative
sentences in FOPL
Then logical reasoning methods are used to deduce consequences
Another name for this type of approach is called “the knowledge-
based” approach
The Symbol Processing Approach uses “top-down” design of
intelligent behavior.
87. Sub-symbolic Approach
Based on the Physical Grounding Hypothesis
“bottom-up” style
Starting at the lowest layers and working upward.
In the sub-symbolic approach signals are generally used rather
than symbols
Proponents believe that the development of machine intelligence
must follow many of the same evolutionary steps.
Sub-symbolic approaches rely primarily on interaction between
machine and environment. This interaction produces and emergent
behavior (evolutionary robotics, Nordin, Lund)
Some other sub-symbolic approaches are: Evolutionary
Computation, Artificial Immune Systems, and Neural Networks
88. Turing Test for Intelligence
Tests the ability of a computer system to act humanly
The aim is to determine if the human interrogator
thinks he/she is communicating with a human.
Turing (1950) "Computing machinery and
intelligence“
"Can machines think?" "Can machines behave
intelligently?“
To pass Turing Test the computer must:
Process natural language;
Represent knowledge;
Reason;
Learn and adapt to the new situations.
Total Turing test included vision & robotics.
90. Modelling an AI System
A typical AI system consists of three
subsystems, i.e.,
Perception Subsystem
Reasoning Subsystem
Action Subsystem(made of actuators/effectors)
92. AI application areas
Game Playing
Much of the early research in state space search was
done using common board games such as checkers,
chess, and the 15-puzzle
Games can generate extremely large search spaces.
These are large and complex enough to require
powerful techniques for determining what alternative
to explore
93. AI application areas
Automated reasoning and Theorem Proving
Theorem-proving is one of the most fruitful branches
of the field
Theorem-proving research was responsible in
formalizing search algorithms and developing formal
representation languages such as predicate calculus
and the logic programming language
94. AI application areas
Expert System
One major insight gained from early work in problem
solving was the importance of domain-specific
knowledge
Expert knowledge is a combination of a theoretical
understanding of the problem and a collection of
heuristic problem-solving rules
Current deficiencies:
Lack of flexibility; if human cannot answer a question
immediately, he can return to an examination of first
principle and come up with something
Inability to provide deep explanations
Little learning from experience
95. AI application areas
Natural Language Understanding and
Semantics
One of the long-standing goals of AI is the
creation of programming that are capable of
understanding and generating human language
96. AI application areas
Modeling Human Performance
Capture the human mind (knowledge
representation)
97. AI application areas
Robotics
A robot that blindly performs a sequence of
actions without responding to changes or being
able to detect and correct errors could hardly be
considered intelligent
It should have some degree of sensors and
algorithms to guide it
98. AI application areas
Machine Learning
Learning has remained a challenging area in AI
An expert system may perform extensive and
costly computation to solve a problem; unlike
human, it usually doesn’t remember the solution
Examples include:
Decision tree learning
Genetic algorithms
Neural networks
99. Application Domains of AI
Application domain areas include:
Military
Medicine
Industry
Entertainment
Education
Business
100. Summary of Today’s Lecture
Intelligence and types
Artificial Intelligence involves the study of:
automated recognition and understanding of signals
reasoning, planning, and decision-making
learning and adaptation
AI has made substantial progress in
recognition and learning
some planning and reasoning problems
…but many open research problems
AI Applications
improvements in hardware and algorithms => AI applications in industry, finance, medicine,
and science.
Rational agent view of AI
Two views of AI: symbol based vs. sub-symbolic AI
Turing test for intelligence and AI applications