TDT4136 Introduction to Artificial Intelligence

Lecture 1: Introduction
(Chapter 1 in the textbook)

Pinar Öztürk

Norwegian University of Science and Technology

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 1 / 41

Practical information

All practical Information is in the first part (separate) of Lecture 1 slides.

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 2 / 41

Course topic overview

Intelligent agents
Problem solving by Searching
Adversarial Search
Constraint Satisfaction Problems
Logical systems
Knowledge representation
Planning
Game Theory
Ethical issues in AI

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 3 / 41

Todays outline

What is AI?
A brief history
The state of the art

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 4 / 41

What is AI?

There are no crisp definitions. Here is one from John McCarthy, (Father of
the phrase Artificial Intelligence)
see http://www.formal.Stanford.EDU/jmc/whatisai/
Question: What is artificial intelligence?
Answer: It is the science and engineering of making intelligent machines,
especially intelligent computer programs. It is related to the similar task of
using computers to understand human intelligence, but AI does not have to
confine itself to methods that are biologically observable.
Question: Yes, but what is intelligence?
Answer: Intelligence is the computational part of the ability to achieve
goals in the world. Varying kinds and degrees of intelligence occur in
people, many animals and some machines.

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 5 / 41

What is AI? - cont

"It is the science and engineering of making intelligent machines, especially

intelligent computer programs" John McCarthy

"AI is the science of making machines to do things that would require

intelligence if done by men." Marvin Minsky

"The study of mental faculties through the use of computational models"

Eugen Charniak

In short: There is no formal definition covering all aspects of intelligence

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 6 / 41

Two aspects of AI

Science
The science of understanding intelligent entitites, developing theories which
attempt to explain and predict the nature of such entities
Discover ideas about knowledge that help explain various sorts of
intelligence
Model functions of the human brain

Engineering
Solving real-world problems by employing ideas of how to represent and use
knowledge
Engineering of intelligent entities
Produce intelligent behaviour by any means

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 7 / 41

Can machines be Intelligent?

Symbolic system hypothesis (Newell and Simon)

- Intelligence is substrate neutral
- A physical symbol system has necessary and sufficient means for
general intelligent action.

Biological substrate only (John Searle, philosopher)

- Intelligence is substrate dependent.
- The material humans are made of is fundamental for our intelligence.
- Thinking is possible only in special machines - living ones made of
proteins.

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 8 / 41

AI prehistory - grounding disciplines

Computer Science is the main discipline underlying AI

Can we think which other disciplines AI is grounded on?

Computer
Science

ArtificiaI
?
Inteligence

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 9 / 41

AI prehistory - grounding disciplines

Computer
science

Psychology Philosophy

ArtificiaI
Control theory Mathematics
& Cybernetics Inteligence

Linguistics

Economics

Neuroscience

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 10 / 41

AI prehistory

Philosophy logic, methods of reasoning

mind as physical system
foundations of learning, language, rationality
Mathematics logic, formal representation
algorithms, computation, probability theory
Statistics modeling uncertainty, learning from data
Psychology phenomena of perception and motor control
cognitive psychology

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 11 / 41

AI prehistory II

Economics formal theory of rational decisions, utility

Linguistics knowledge representation
grammar, syntax, semantics
Neuroscience neurons as information processing units
synapse as learning mechanism
Control theory homeostatic systems, stability
simple optimal agent designs, maximize objective function
Computer science engineering, hardware
computational complexity theory

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 12 / 41

Birth of AI field

A Summer Research Project in the Dartmouth College, in 1956 was the birth of the AI
research field.

(From left: Trenchard More, John McCarthy, Marvin Minsky, Oliver Selfridge, and Ray
Solomonoff)

McCarth: "...to proceed on the basis of the conjecture that every aspect of learning or
any other feature of intelligence can in principle be so precisely described that a machine
can be made to simulate it."

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 13 / 41

Brief history of AI

1943 McCulloch & Pitts: Boolean circuit model of brain

1950 Turing’s “Computing Machinery and Intelligence”
1952-69 Early AI programs, including Samuel’s checkers program,
Newell & Simon’s Logic Theorist, Gelernter’s Geometry Engine
1956 Dartmouth meeting: “Artificial Intelligence” adopted
1962 Rosenblatt’s Perceptron for training simple neural networks
1965 Robinson’s complete algorithm for logical reasoning
1972 The logic programming language PROLOG is created.
1966–74 Disappointment: AI discovers computational complexity
Neural network research almost disappears

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 14 / 41

Brief history of AI -cont.

1969–79 Early development of knowledge-based systems

1980–88 Expert systems industry booms
1988–93 Expert systems industry busts: “AI Winter”
1985–95 Backpropagation learning returns neural networks to popularity
1988– Resurgence of probability; general increase in technical depth
“Nouvelle AI”: ALife, GA
1995– Agents, agents, everywhere . . .
2003– Human-level AI back on the agenda
2005-2010 2005-2010 AI disappoints again, AI is not much appreciated

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 15 / 41

Example from Early History of AI
1943 McCulloch & Pitts: Boolean circuit model of brain

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 16 / 41

McCulloch and Pitts 1943

Goal: to understand how the brain produces complex thoughts

("propositions") by using (simple) neurons
- model of a neuron, axon, dendrode - "MCP neuron"
- network of neurons
- transfer of information through on/off mechanism of neurons
- every network of MCP neurons encodes some logical proposition

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 17 / 41

McCulloch and Pitts 1943 - "Bird example"
Example from:
http://www.mind.ilstu.edu/curriculum/mcp_neurons/mcp_neuron_
1.php?modGUI=212&compGUI=1749&itemGUI=3018

Assume a bird will "decide" to eat/not an object (e.g., blue berry, orange,
basketball, daisy). It receives/perceives two inputs about the object:
- shape: round or not
- colour: purple or not
- decision: eat only if round purple object (e.g., blue berry)

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 18 / 41

Bird example - cont.

- two inputs (each takes value 0 or 1)

- threshold T
- output is a function of inputs and T
- IF(sum(inputs))>= T
upon (1) the information it receives from the environment (through it's two sensory
"detectors"), and (2) the threshold at which the MCP neuron is set. The detectors can

THEN Output=1
be thought of as "looking for" particular properties. The combined input to the neuron
(from the detectors) will be larger or smaller depending upon how many of the
"looked-for" properties are detected by the sensors. Whether or not the total input
from the sensors will be a number high enough to cause the neuron to "fire" (i.e, to
give an output of "1" instead of "0") depends upon how high the threshold is.

We can represent all of this information in a table. Table 3 (below), for example, shows
the situation if the threshold is set at 1. You will see that this would cause the bird to
eat not only food that is proper (blueberries) but also food that it should not eat
(golfballs and violets).

Table 3

Total Greater
Object Purple? Round? than or Eat?
equal to
threshold
of 1?

Blueberry 1 1 2 Yes 1

Golf ball 0 1 1 Yes 1

Violet 1 0 1 Yes 1

Hot Dog 0 0 0 No 0

These are the basic characteristics of a simple MCP neuron that takes only positive
numbers as input. Let us now give the neuron slightly more flexibility by adding
negative input values.

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 19 / 41

Some killer apps in AI

1991 During the Gulf War, US forces deployed an AI logistics planning and
scheduling program that involved up to 50,000 vehicles, cargo, and people.
Saved the US more money than spent on all AI research since 1950
1997 Deep Blue (IBM) defeated world chess champion Gerry Kasparov

2011 Watson (IBM) beat human champions on “Jeopardy”

2012 Google car obtains driver’s license in Nevada, US. By 2014, the cars
have driven for 1.1 million km without accident

2017 DeepMind’s Alphago AND Elon Musk’s A.I. Destroys Champion

Gamer! https://www.youtube.com/watch?v=XbDmxEOj9OY

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 20 / 41

What is AI?

Russell and Norvig’s definition of AI

Two dimensions:

AI approaches
criteria

process/task evaluation

Performance Performance
Thinking acting
evaluated evaluated
in comparison according to
with human rationality

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 21 / 41

What is AI?

Types/approaches to AI - as defined by Russel and Norvig:

humanly rational

think

act

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 22 / 41

What is AI? -cont

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 23 / 41

Thinking humanly: Cognitive Science

1960s “cognitive revolution”: information-processing psychology replaced

prevailing orthodoxy of behaviorism.
Tries to form computational theories of internal activities of the brain.
How to validate? Requires
1 Predicting and testing behavior of human subjects (top-down), or
2 Direct identification from neurological data (bottom-up)
Both approaches (roughly, Cognitive Science andCognitive
Neuroscience) are now distinct from AI

Both share with AI the following characteristic:

the available theories do not explain (or engender)
anything resembling human-level general intelligence

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 24 / 41

John McCarthy talks about AI

https://www.youtube.com/watch?v=Ozipf13jRr4

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 25 / 41

Acting humanly: The Turing test
Turing (1950) “Computing machinery and intelligence”:
Can machines think?” −→ “Can machines behave intelligently?”
Operational test for intelligent behavior: the Imitation Game

HUMAN

HUMAN
INTERROGATOR ?
AI SYSTEM

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 26 / 41

Acting Humanly - Eliza

One of the most famous early AI programs: Eliza, the computer

psychotherapist, created by Joseph Weizenbaum in 1966 at MIT.
Eliza functions by "twisting the statements of her ’patients’ back at
them in the classic manner of a non-directive psychotherapist."
The fact that it understands the subject’s statements is an illusion.
Surprisingly, many users were taking its performance quite seriously.

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 27 / 41

Acting Humanly - Eliza 2
young woman: Men are all alike.
eliza: In what way?
young woman: They're always bugging us about something
specific or other.
eliza: Can you think of a specific example?
young woman: Well, my boyfriend made me come here.
eliza: Your boyfriend made you come here?
young woman: He says I'm depressed much of the time.
eliza: I am sorry to hear you are depressed.
young woman: It's true. I'm unhappy.
eliza: Do you think coming here will help you not to be unhappy?
young woman: I need some help, that much seems certain.
eliza: What would it mean to you if you got some help?
young woman: Perhaps I could learn to get along with my mother.
eliza: Tell me more about your family.
young woman: My mother takes care of me.
eliza: Who else in your family takes care of you?
young woman: My father.
eliza: Your father.
young woman: You are like my father in some ways.
eliza: What resemblance do you see?
young woman: You are not very aggressive but I think you don't
want me to notice that.
eliza: What makes you think I am not very aggressive?
young woman: You don't argue with me.
eliza: Why do you think I don't argue with you?
young woman: You are afraid of me.
eliza: Does it please you to believe I am afraid of you?
young woman: My father is afraid of everybody.
eliza: What else comes to mind when you think of your father?
young woman: Bullies.
eliza: Does that have anything to do with the fact that your
Pinar made
boyfriend Öztürk (NTNU)
you come here? TDT4136 Introduction to Artificial Intelligence 28 / 41
How humany is Sofia (Hong Kong Kong firm Hanson
Robotics)?

https://www.youtube.com/watch?v=suRuQbDXcrc

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 29 / 41

Thinking rationally: Laws of Thought-1

Aristotle: what are correct arguments/thought processes?

Formalize “correct” reasoning using a mathematical model
Several Greek schools developed various forms of logic:
notation and rules of derivation for thoughts;
Direct line through mathematics and philosophy to modern AI

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 30 / 41

Thinking rationally: Laws of Thought-2

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 31 / 41

Thinking rationally: Laws of Thought-3

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 32 / 41

Thinking rationally: Laws of Thought -4

Problems:
1) Not all intelligent behavior is mediated by logical deliberation
2) What is the purpose of thinking? What thoughts should I have
out of all the thoughts (logical or otherwise) that I could have?
3) Formalizing (informal) common sense knowledge is difficult
4) General deductive inference is computationally intractable

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 33 / 41

Acting rationally: Rational Agents approach

Rational behavior: doing the right thing

The right thing: that which is expected to maximize goal achievement,
given the available information and computational abilities
Doesn’t necessarily involve thinking—e.g., blinking reflex—but
thinking should be in the service of rational action

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 34 / 41

Two main AI paradigms

Good old fashioned AI (GOFAI)

Situated embodied AI (SEAI)

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 35 / 41

Good old fashioned AI

Most of the successes of AI in the 1970s and 1980s were due to

research based on Newell’s and Simon’s Physical Symbol System
Hypothesis
A physical symbol system has the necessary and sufficient means for
general intelligent action.
Newell and Simon viewed intelligence as symbol manipulation, and
hypothesized that it didn’t make difference what physical medium –
brain, paper, or computer – was used to do the symbol manipulation
Hence a special emphasis on symbolic representations, which can be
interpreted as representing situations in the real world, e.g. "a block
world"

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 36 / 41

Critiques of GOFAI

Searle: a program (or any physical symbol system) could not be said
to understand the symbols that it uses; the symbols have no meaning
for the machine
Brooks: our most basic skills of motion, survival, perception, balance
etc. do not seem to require high level symbols at all; the use of high
level symbols was more complicated and less successful
Harnad: the symbol grounding problem: an agent does not perceive
symbols, instead the brain converts sensory inputs into higher level
abstractions, e.g. symbols

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 37 / 41

Emergence of SEAI

The GOFAI approaches turned out to be brittle and very little robust
when deployed on real-world problems
Trying to define a model of the world turned out to be quite hard -
this led to Brooks’ statement that “the world is its own best model”
Situated and embodied AI focuses on having a body (i.e. motor skills)
in a physical environment
Swarm intelligence, subsymbolic AI, genetic algorithms, neural
networks - however, this is not a big part of this course, this is covered
in more detail in other AI courses.

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 38 / 41

Course topic overview

Next Lecture: Intelligent Agents (chapter 2 in R&N book)

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 39 / 41

More Videos

Computers with common sense( Doug Lenat):

https://www.youtube.com/watch?v=2w_ekB08ohU

Artificial Intelligence: The Common Sense Problem (Hubert Dreyfus)

https://www.youtube.com/watch?v=SUZUbYCBtGI

AI Boom (Rodney Brooks):

https://www.youtube.com/watch?v=ovbzjLobiDo

IBM Watson : https://www.youtube.com/watch?v=_Xcmh1LQB9I

Jeopardy : https://www.youtube.com/watch?v=WFR3lOm_xhE

State of the art AI (Minsky):

https://www.youtube.com/watch?v=aODnFdU_hds

Pinar Öztürk (NTNU) TDT4136 Introduction to Artificial Intelligence 40 / 41