Fundamentals of

Artificial
Intelligence
Debre Berhan University
February 2016E.C
1
TABLE OF CONTENT
• Intelligence • History of AI

• Definition of AI • Application of AI

• What is AI? • State-of-art of AI

• Types of AI • Advantage and

• Foundation of AI Disadvantage of AI

2
Intelligence?
“Intelligence: T h e ability t o learn and solve problems”
Webster’s Dictionary.

“Intelligence: capacity of abstraction, learning, logic,

understanding, self-awareness, emotional knowledge,
reasoning, planning, critical thinking and problem
solving.
Wikipedia.

“Intelligence is multifaceted mental ability that involves

to learn, understand, plan,, solve problem and adapt to
new situations.
Chat GPT.

3
 Definition of AI
“Artificial intelligence ( A I ) is t h e intelligence exhibited
by machines or software’
Wikipedia.

“ T h e science and engineering of making intelligent

machines”
McCarthy.

“ T h e study and design of intelligent agents, w here an

intelligent agent is a system t h a t perceives its environment
and takes actions t h a t maximize its chances of success.”
Russel and Norvig AI book.
4
Definition of AI

“AI is the development of computer Systems

that can perform tasks that typically requires
human intelligence such as perception, speech
recognition,, decision making and translation”

5
Why AI?
With the help of AI,
• we can create software/devices which can solve real-world problems
such as health issues, marketing, traffic issues, etc efficiently and
accurately.
• we can create a personal virtual Assistant, such as Cortana, Google
Assistant, Siri, etc.
• we can build Robots which can work in an environment where survival
of humans can be at risk.
AI opens a path for other new technologies, new devices, and new
Opportunities .

“Just as t h e Industrial Revolution freed up a lot of humanity f r o m

physical drudgery, I think A I has t h e potential t o free up humanity
f r o m a lot of t h e m e n t a l work.”
Andrew Ng. 6
Goal/Objective of AI
Replicate human intelligence

Solve Knowledge-intensive tasks

 An intelligent connection of perception and action

Building a machine which can perform tasks that requires

human intelligence such as:
 Proving a theorem, Playing chess, Plan some surgical operation and
Driving a car in traffic

7
 Types of AI
General Vs Specific AI
General or Strong AI
• Aims to build machines that can truly reason and solve problems. These machines
should be self-aware and their overall intellectual ability needs to be
indistinguishable from that of a human being.
• Strong AI maintains that suitably programmed machines are capable of cognitive
mental states.

Specific/Narrow or Weak AI
• Deals with the creation of some form of computer-based AI that cannot truly reason
and solve problems, but can act as if it were intelligent.
• Weak AI holds that suitably programmed machines can simulate human cognition.

8
 Cont.… Types of AI
Other AI Types

Applied AI
• Aims to produce commercially viable "smart" systems such as, a security system that
is able to recognize face.
• Applied AI has enjoyed considerable success.

Cognitive AI
• Computers are used to test theories about how the human mind works--for
example, theories about how we recognize faces and other objects.

9
What is AI?
T h e r e ar e f our d i f f e r e n t a p p r o a c h e s / schools of thoughts;
all have been followed by different people w ith different
methods. (Russel & Norvig)

10
 What is AI?
I: Thinking humanly: cognitive modelling approach

 Requires to determine how humans think! ……..1960’s “cognitive

revolution”.
 Requires scientific theories of internal activities of the brain
 Through introspection: trying to catch the thoughts as they go by
 Through Psychological Experiments: observe the person in action
 Through brain imaging: observe brain in action

 The Assumption being; Once we have the precise theory of the mind, it is possible to
express the theories of in computer program.

Today, Cognitive Science and AI are distinct disciplines.

11
 What is AI?
II: Acting humanly: The Turing Test approach
Turing test ( Al a n Turing 1 9 5 0 ) : A computer passes the test of
intelligence, if it can fool a human interrogator.

The computer would need to possess the following capabilities:

1. Natural language processing
2. Knowledge representation
3. Automated Reasoning
4. Machine Learning

12
 What is AI?
II: Acting humanly: The Turing Test approach

T h e To t a l Tu r i n g t e s t i n c l u d e s v i d e o s i g n a l s s o t h a t i n t e r r o g a t o r
c a n t e s t t h e s u b j e c t s p e r c e p t u a l a b i l i t y. To P a s s t o t a l Tu r i n g
Te s t t h e c o m p u t e r n e e d s
1 . C o m p u t e r Vi s i o n a n d
2. Robotics

N ot e : The s e Si x D i s c i p l i n e s C o m p o s e s M o s t o f A I .

Acting humanly?

13
 What is AI?
III: Thinking rationally: the “law of thought” approach
• This law of thoughts were supposed to govern the operation of mind; which initiated
the field of logic.

• Codify “right thinking” with logic. Given correct inferences can lead to correct
conclusion.

• Example: Socrates is Man; All men are mortal; Therefore Socrates is Mortal.

• Several Greek schools developed various forms of logic:notation and rules of

derivation for thoughts.

Problems/Obstacles of this Approach:

1. Not all knowledge can be expressed with logical notations.
2. Computational blow up. There is also a difference In solving problem in principle and in
practice.
14
 What is AI?
IV: Acting rationally: the rational agent approach
 Agent is something that acts;
 A rational agent is one that acts so as to achieve the best outcome, or when there is
uncertainty, the best expected out- come.
 “the law of thoughts” emphasis on correct inference which is part of being rational agent but
not always.

 All the skills needs for Turing test also allows an agent to act rationally.

 Rational Agent Advantage over other Approaches:

1. It is more general
2. It is more amenable to scientific development that approaches based on human
behavior. Rationality is mathematically well defined and completely general.

Thus , o u t A I d e f i n i t i o n [ A p p r o a c h e s ] c o n c e n t r a t e s o n g e n e r a l p r i n c i p l e s
of r a t i ona l a ge nt a n d o n c o m p o n e n t s f o r c o n s t r u c t i n g t h e m ! !

15
The Foundation of AI
 Rationalism, Dualism, Materialism,
Philosophy  Logic, methods of reasoning.
 Mind as physical system that operates as a set of rules.
 Foundations of learning, language, rationality.
 Logic: Formal representation and proof.
Mathematics  Computation, algorithms.
 Probability.

 Economics; people make choices that lead to preferred outcomes.

Economics  Formal theory of rational decisions.
 Combined decision theory & probability theory for decision making under uncertainty.
 Game theory.
 Markov decision processes.

Neuroscience – Study of brain functioning.

– How brains and computers are (dis)similar.

16
Cont... The Foundation of AI
 How can artifacts operate under their own control?
Control theory  Design simple optimal agents receiving feedback from the environment.
and cybernetics  Moderncontrol theory design systems that maximize an objective function over time.

Computer • Cares about how to build powerful machines to make AI possible.

Engineering • E.g., Self-driving cars are possible today thanks to advances in computer
engineering.
– How do we think and act?
Psychology – Cognitive psychology perceives the brain as an information processing machine.
– Led to the development of the field cognitive science: how could computer
models be used to study language, memory, and thinking from a psychological
perspective.

Linguistics – How are language and thinking related?

– Modern linguistics + AI = Computational linguistics (Natural
language processing).

17
 History of AI
• Gestation of AI (1943-1955):
• First work recognized as AI was done by McCulloch & Pitts:
• They drew on three sources: knowledge physiology and function of neurons in the
brain; propositional logic; and Turing’s theory of computation.
• They proposed a model of artificial neurons;
• Turing’s Computing Machinery and Intelligence.
• The birth of AI (1956):
• Birth of AI @ Dartmouth 2-month Workshop 1956.
• McCarthy with Minsky, Claude Shannon, and Nathaniel Rochester bring together U.S.
researchers interested in automata theory, neural nets, and the study of intelligence.
• There were 10 attendees in all, including Trenchard More from Princeton, Arthur
Samuel from IBM, and Ray Solomonoff and Oliver Selfridge from MIT.
18
 History of AI
• Early enthusiasm, great expectations (1952-1969):
– Early years of AI were full of successes.
– GPS was probably the first program to embody the “thinking humanly” approach.
– Early work building on the neural networks, Hebb’s learning methods were enhanced, concepts of
perceptrons introduced.
• The dose for reality ( 1966-1973):
• these early success fail miserably when tried out on wider selections and more difficult problems.
• 1st difficulty is that early success were by means of simple syntactic manipulations. Machine
translation efforts to speed up the translation of Russian scientific papers in the wake of the Sputnik
launch in 1957 has failed; all U.S. government funding for academic translation projects was
canceled.
• The second difficulty was the intractability of many of problems that AI was attempting to solve.
The fact that a program can find a solution in principle does not mean that the program contains
any of the mechanisms needed to find it in practice.
• Thirdly; some fundamental limitations on the basic structures being used to generate intelligent
behavior…. Example: perceptrons (a simple form of neural network) 19
 History of AI
• Knowledge-based Systems (1969-1979):
• the first successful knowledge-intensive system.
• applications to real-world problems increase the demand.
• AI Becomes an Industry (1980 – Present)
• The first successful commercial expert system, R1
• AI industry boomed from a few million dollars in 1980 to billions of dollars in 1988,
• Soon “AI winter” came; the period where companies failed to deliver on extravagant
promises.

20
 History of AI
• The return of Neural networks (1986- Present)
• different groups reinvented the back-propagation learning algorithm first found
in 1969
• AI adopts the Scientific method (1987-present)
• Methodology; AI has finally come firmly under the scientific method.
• Speech Recognition, Hidden Markov Models, Machine Translation., Bayesian Networks
• The emergence of Intelligent Agents (1995 – Present)
• The “whole agent” approach
• The Availability of very large data sets (2001 – Present)
• More emphasis for the data rather than the algorithm
• Data will drive future discoveries and alleviate the complexity in AI.
• AI Spring! 21
State-of-the-art of AI Applications
• Speech recognition • Fraud detection
• Autonomous planning scheduling • Recommendation systems
• Financial forecasting • Web search engines
• Game playing, video games • Autonomous cars
• Spam fighting • Energy optimization
• Logistics planning • Question answering systems
• Robotics (household, surgery, • Social network analysis
navigation) • Medical diagnosis, imaging
• Machine translation • Route finding
• Information extraction • Traveling salesperson
• VLSI layout • Protein design
• Automatic assembly • Document summarization
• Sentiment analysis • Transportation/scheduling
• Computer animation
22
Application Handwriting recognition
(check, zipcode)
Machine translation

of AI
• Historical motivation: translate Russian to
English.

• First systems using mechanical translation

(one-to-one cor-respondence) failed!
• “Out of sight, out of mind” ) “Invisible,
imbecile”.
• Today, Statistical Machin Translation
leverages the vast amounts of available
Robotics:Awesome robots today! translated corpuses.
NAO, ASIMO, and more! • While there is room for improvement, machine
translation has made significant progress.

23
 Cont…
Recommendation systems
Search Engines
(collaborative filtering)
• Google search Engine;
[YouTube, amazon,
Application of AI Facebook recommender • Bing, Yahoo search engines
systems ]

Face Recognition
Face Detection

24
 Cont… Others
• Natural Language Understanding
Application of AI and information extraction!

• Google AlphaGo: Deep

Learning, reinforcement learning,
and search algorithms!

Detection of breast cancer in • Autonomous driving

mammography images
• Chess ( 1 9 9 7 ) : Kasparov vs.
I B M Deep Blue

25
Advantage and Disadvantage of AI
Advantage Disadvantage
High Accuracy with fewer errors  High Cost
 High-Speed  Can't think out of the box

 High reliability  No feelings and emotions

 Useful for risky areas  Increase dependence on machines

 Digital Assistant  No Original Creativity

 Useful as a public utility

26
AI Founders
• Aristotle • Herbert Simon
• Alan Turing • Allen Newell
• John McCarthy • David Waltz
• Warren McCulloh • Tom Mitchell
• Walter Pitts • Stuart J. Russell
• Claude Shannon • Peter Norvig
• Marvin Minsky • etc.
• Dean Edmonds

27
Thank You!!

28