“ Fundamentals of Artificial Intelligence ”

Instructor: Melaku M.
Target Group: G4 Software Engineering
1
 Chapter 1

Introduction to Artificial Intelligence

Melaku M.

2
Target Group: G4 Software Engineering
1/24/2024 2
 Why the interest in AI?

Medicine/
Diagnosis

1/24/2024 3
 Why the interest in AI?

1/24/2024 4
 What is Artificial Intelligence?
AI is the science and engineering of making intelligent machines,
especially computer programs that are capable of simulating human
intelligent behavior.

❖What is intelligence then? ▪What’s involved in Intelligence?

▪Intelligence is the mental quality. It

is the ability to learn , to solve
problems, and to adapt to new
situations.
 Cont.…

➢Artificial Intelligence is the branch of computer science dedicated on

the development of computer systems or machines capable of
performing tasks that would typically require human intelligence.
➢Examples of these intellectual tasks are:-
▪ Speech recognition, language translation
▪ Performing complex surgical operation
▪ Driving a car in traffic( crowded env’t)
▪ Writing poetry
▪ Proving a mathematical theorem
6
 Cont.…

➢Researchers have pursued several different versions of AI. Some have defined
intelligence in terms of fidelity to human performance, while others prefer an
abstract, formal definition of intelligence called rationality—loosely speaking,
doing the “right thing.” The subject matter itself also varies: some consider
intelligence to be a property of internal thought processes and reasoning, while
others focus on intelligent behavior, an external characterization.
➢From these two dimensions—human vs. rational and thought vs. behavior—
there are four possible combinations, and there have been adherents and research
programs for all four.

7
Approaches to AI:- Views of AI fall into four categories
 Acting Humanly: The Turing Test
Alan Turing(1950) published "Computing
machinery and intelligence":
"Can machines think?" → "Can machines
behave intelligently?“
He has proposed an operational test called:
Imitation game. Answers

Approaches of Turing Tests:-

➢ Interrogator asks the question, both computer & Answers

human via teletype comm., and receives the

answer from both of them. Human
➢ Note: all three entities sit in an isolated location.
➢ Computer passes the test if the interrogator Turing test: ultimate test for testing
cannot distinguish the answers responded by
intelligent behavior of machines
9
computer and humans.
 Acting Humanly: The Turing Test
Captcha: Completely Automated Public Turing test to tell
Computers and Humans Apart.

1/24/2024 10
 Cont.…
►To pass classical Turing Test computer needs to have the following
capabilities (Skills Required)
➢Natural language processing: Enable computers to communicate in
human language.
➢Knowledge representation: Schemes to store information effectively
& efficiently.
➢Automated reasoning: To answer questions & to draw new
conclusions.
➢Machine learning: To adapt to new circumstances and to detect and
extrapolate patterns.
 Cont.…

►Turing viewed the physical simulation of a person as unnecessary to

demonstrate intelligence. However, other researchers have proposed a
total Turing test, which requires interaction with objects and people in
the real world.

►To pass the total Turing Test, the computer will need

1. Computer vision: To perceive and recognize objects.

2. Robotics: To take an actions or modify its environment.

 Thinking Humanly: Cognitive modeling approach

❖To say that a program thinks like a human, we must know “how humans
think”. Requires the knowledge of brain functions(internal activities of the
brain)

❖We can learn about human thought in three ways:

✓Introspection—trying to catch our own thoughts as they go by(looking inward);
✓Psychological experiments—observing a person in action;
✓Brain imaging—observing the brain in action.

❖Once we have a sufficiently precise theory of the mind, it becomes

possible to express the theory as a computer program.
 Thinking Humanly: Cognitive modeling approach

❖For example, Allen Newell and Herbert Simon, who developed AI program
GPS, the “General Problem Solver” were not content merely to have their
program solve wide ranges of problems correctly. They were more
concerned with comparing the sequence and timing of its reasoning steps to
those of human subjects solving the same problems.

❖Cognitive Science: interdisciplinary field that brings together computer

models from AI and experimental techniques from psychology to construct
the working of the human mind.
 Thinking rationally: Laws of thought approach
➢The Greek philosopher Aristotle was one of the first to attempt to codify “right
thinking”— that is, irrefutable reasoning processes.
➢His syllogism provided the patterns for argument structures that always yielded correct
conclusions when given correct premises.
➢For example, “Socrates is a man; all men are mortal; therefore Socrates is mortal.‘’
➢These laws of thought were supposed to govern the operation of the mind; their study
initiated the field called logic.
➢Logicians developed various forms of logic notation and rules of derivation for
thoughts;
➢Problems to this approach;
▪ Logic as conventionally understood requires knowledge of the world that is certain.

▪ Not all intelligent behavior is controlled/mediated by logic.

 Acting Rationally: Rational agent approach
►Rational behavior: doing the right thing.

►The right thing: that which is expected to maximize goal achievement, given the
available information.

►In the “laws of thought” approach to AI, the emphasis was on correct inferences.
Making correct inferences is sometimes part of being a rational agent, because one
way to act rationally is to deduce that a given action is best and then to act on that
conclusion.

►On the other hand, there are ways of acting rationally that cannot be said to
involve inference/thinking—e.g., blinking, recoiling from a hot stove reflex
16
actions.
 Rational Agent
❖AI has focused on the study and construction of agents that do the
right thing.
❖For any given class of environments and tasks, we seek the agent (or
class of agents) with the best performance.
❖A rational agent is one that acts so as to achieve the best outcome or,
when there is uncertainty, the best expected outcome.
❖Caveat: computational limitations make perfect rationality
unachievable.
→ design best program for given machine resources 17
 Goals of AI
Intelligent Problem-Solving and Decision-Making: At its core, AI aims to
simulate human intelligence in machines, empowering them to carry out complex
tasks and decision-making processes autonomously.

To Create Expert Systems −an intelligent computer program that solves the
problems in a specialized domain that normally requires human expertise.

To make computers more useful by letting them take over dangerous and
complex tasks from human.

• To reduce human causalities in war, dangerous workplaces, natural disaster…

To solve knowledge intensive task
 Goals of AI
Expanding Human Capabilities: AI can assist humans in tasks beyond our
physical or cognitive limitations. For example, AI algorithms can analyze data too
complex for human comprehension.
Advancing Scientific Discovery and Innovation: AI can accelerate scientific
progress by analyzing vast amounts of data, identifying new patterns, and
suggesting hypotheses. AI Fostering Creativity and Innovation.
Human-Like Learning and Adaptation: Ideally, AI systems should learn and
adapt from experience like humans. This could involve continual improvement
through interactions with their environment, allowing them to handle
increasingly complex tasks and unexpected situations.
NLP: AI strives to develop systems that can understand spoken and written
language nuances, in much the same way human being can.
 Disciplines that contributed ideas, viewpoints and techniques to AI:
• AI has identifiable roots in a number of disciplines, particularly:
• Philosophy
• Can formal rules be used to draw valid conclusions?
• Aristotle formulated a precise set of laws(syllogism) governing the rational part of the mind.
• Where does knowledge come from?
• How does knowledge lead to action? Can computers have ”consciousness”?
• Mathematics
• What are the formal rules to draw valid conclusions?
• Proof, computation, (un)decidability, (in)tractability.
• How do we reason with uncertain information?
• Probability helping to deal with uncertain measurements and incomplete theories

• Control theory
• How can artifacts operate under their own control?
• Design systems that maximize an objective function over time
 Disciplines that contributed ideas, viewpoints and techniques to AI:

• Psychology /Cognitive psychology

• How do humans and animals think and act?
• Understanding, how the human mind works/functions, and lead to intelligent
behavior. Cognitive psychology initiated
• Computer science and Engineering
• How can we build an efficient computer? Design efficient algorithms and graphics
• Economics
• Utility, decision theory, rational economic, agents maximizing their own well being.
• Linguistics
• For understanding natural languages different approaches has been adopted from
the linguistic work. Modern linguistics and AI were “born” computational
linguistics or NLP.
 History of AI
The following are some of milestones in the history of AI which defines the journey from
the AI generation to till date development. How AI is progressed /evolved at different year?

22
 Cont.…

A. Maturation of Artificial Intelligence (1943-1952) B. The birth of Artificial Intelligence (1952-1956)

✓ Warren McCulloch & Walter Pitts (1943) : ✓ Allen Newell and Herbert A.
proposed a model of artificial neurons. Simon(1955) : created first AI program,

✓ Donald Hebb (1949) : demonstrated an named Logic Theorist. This program

updating rule for modifying the connection proved 38 of 52 mathematics theorems.

strength between neurons. His rule is called ✓ John McCarthy(1956) : The word
Hebbian learning. "Artificial Intelligence" first time

✓ Alan Turing (1950) : He pioneered Machine adopted and called as academic field.

learning & published “computing machinery Also, high-level computer languages

& intelligence” in which he proposed test such as FORTRAN, LISP, or COBOL

(i.e. can machines think?). were invented at that time.

23
 Cont.…
C. Golden years-Early enthusiasm (1956-1974) E. A boom of AI (1980-1987)
✓Joseph Weizenbaum(1966): created the ✓After AI winter duration AI came back with
first chatbot which is called ELIZA. "Expert System".
✓In the year 1972 : The first intelligent F. The Second AI winter (1987-1993)
humanoid robot was built in Japan which
✓Again, investors and government stopped in
was named WABOT-1.
funding for AI research due to high cost but
D. The first AI winter (1974-1980) not efficient results.
✓Because, computer scientists dealt with a G. The Emergence of Intelligent Agents (1993-2011)
severe shortage of funding from the ✓In the year 1997, IBM Deep Blue beats
government for AI researches. world chess champion, Gary Kasparov, and
✓Publicity interest on AI was decreased. became the first computer to beat a world chess
24
champion.
 Cont.…
✓In the year 2002, AI entered the home in ✓In the year (2012) : Google has launched an
the form of Roomba, a vacuum cleaner. Android app feature "Google now“. It is one

✓In the year 2006 : AI came into the of voice activated google predictive search

Business world. Companies like app.

Facebook, Twitter, and Netflix also started ✓In the year (2014): Chatbot "Eugene
using AI. . Goostman" won a competition in the

H. Deep Learning, Big data and Artificial infamous "Turing test."

General Intelligence (2011-present) ✓In the year (2018): The "Project Debater"

✓In the year 2006: IBM's Watson won from IBM debated on complex topics with

Jeopardy, a quiz show. Watson had proved two master debaters and also performed

that it could understand natural language extremely well.

and can solve tricky questions quickly. 25

 State of the Art
Intelligent Systems in Your
Everyday Life:
✓Self Driving Cars
✓Drones
✓Chatbots
✓Creating art (such as poetry)
✓Virtual assistants: Alexa, Siri,
Google assistant etc.
✓Navigation and Travel:
suggesting the hotels, flights, and
best routes to the customers.
 State of the Art

►Gaming − AI plays a crucial role in

strategic games such as chess, poker, etc.,
where machine can think of large number
of possible positions based on heuristic
knowledge.
1/24/2024 27
 State of the Art
✓Banks: automatic check readers, signature
verification systems, automated loan
application classification, detect fraudulent
credit card transactions
✓Netflix: Movie recommendation
✓ Google: Predictive search
✓Pattern recognition: handwritten character
or face identification ✓ Autonomous planning & scheduling:
✓Natural Language Processing tasks NASA's on-board program controlled the
(NLP): operations for a spacecraft a hundred
➢Machine Translation, summarization million miles from Earth
✓ Game playing: IBM’s Deep Blue, Google’s
➢Spell Checking, grammar checking
AlphaGo
 State of the Art

❖ Speech recognition has become ❖ (Deep) Question Answering

useful enough to be a convenient ✓ IBM Watson / Jeopardy!
input modality. ✓ Apple Siri
• Most advanced use in personal • “Hey Siri, what’s the best sushi
assistants: Apple’s Siri, Google place in town?”
Assistant, Amazon Alexa, • “Find videos I took at Iva’s birthday
Microsoft Cortana etc. party”
• “What movies are playing today?”
• “Wake me up at 7 AM tomorrow”
1/24/2024 29
 State of the Art
❖In E-Commerce ❖Robotics
• Personalized Shopping: Create • Carrying goods in hospitals, factories,
recommendation engines in accordance
and warehouses, cleaning offices and
with customer browsing history,
Inventory management
preference, and interests.
• AI-Powered Assistants: AI ❖Facial Recognition
chatbots that are used to provide • Used by Facebook to tag photos
customer support/service
• Our phones, laptops, and PCs use facial
• Fraud Prevention: Credit card frauds
recognition techniques in order to
and fake reviews are two of the most
provide secure access.
significant issues that E-Commerce
companies
1/24/2024
deal with. 30
 State of the Art
❖In Education:
• Learning assistance: present simple • Smart Content creation:
yet intuitive explanations for an Developing Course Content
enhanced learning experience. E.g., • Immersive Learning: Immersive
Microsoft Math application. technologies like Augmented reality
• Personalized Learning: Not every (AR) and virtual reality (VR) can
simulate real-world scenarios for
student adapts to knowledge the same
students to gain hands-on experience
way. With personalized learning, each
with low risk, like medical
student can have their own way/style
simulations and technical
of learning as per their level of experiments.
understanding and need.
1/24/2024
• Administrative task automation31
Types of AI

32
 i) Narrow AI

►Narrow AI, also called as Weak AI, focuses on one narrow task and cannot
perform beyond its limitations.
►The most common and currently available AI is Narrow AI in the world of AI
►Narrow AL, includes an AI system that can perform narrowly defined specific
tasks just like a humans or outperform humans.
►Virtual assistants such as Alexa, Siri, Google Assistant are an example of a
Narrow AI that operates with a limited pre-defined range of functions. They
often have problems with tasks outside its breadth of abilities.
►Some examples of Narrow AI are Google translate, playing chess, self-driving
cars, speech recognition, spam filtering and image recognition.

33
 ii) General AI

►General AI is a type of intelligence that could perform any intellectual task

with efficiency like a human.
❖General AI, system have human-level cognitive capabilities in a variety of
disciplines.

►The idea behind the general AI is to make such a system which could be
smarter and think like a human by its own.
►Currently, there is no such AI system exist which can perform any task as
perfect as a human.

►Makes the bold claim that computers can be made to think on a

level (at least) equal to humans 34
 iii) Super AI

►Super AI is a level of Intelligence of Systems at which machines could

surpass human intelligence, and can perform any task better than a human
with cognitive properties.

►Some key characteristics of strong AI capability include the ability to

think, to reason, problem solving, creativity, make judgments, plan, learn,
and communicate by its own.

►Super AI is still a hypothetical concept of Artificial Intelligence.

35
36
1/24/2024 37