Artificial Intelligence (III/II)

Course Code: CT-653

(Module#1)
Dhawa Sang Dong
(Lecturer)

Kathmandu Engineering College

Kalimati, Kathmandu
January 2023
 Chapter#1
Introduction to Artificial Intelligence

✓ Class Outline

1 Introduction to Artificial Intelligence

2 Brief History of Artificial Intelligence

3 Importance and Applications of Artificial Intelligence

4 Knowledge and Learning

5 Intelligent Agent
 Course Evaluation
Theory (100)
I Internal weight (20/100)
- Assignments (in total 6/7) for each module [ 10 ]
- Class Attendance, Activities and 3 - ADT [ 4 + 6 = 10 ]
3 - ADT [ 1 + 2 + 3 = 6 ]
II External weight (80/100)
- End Semester Exam by IOE, TU
Practical (25) – Internal only
- Lab Attendance [ 5 ] – 1 marks per Lab.
- Lab Report and Minor ML Project (Python) [ 10 + 10 = 20 ]

References
- Artificial Intelligence A modern approach – Stuart J. Russell
and Peter Norvig
- Introduction to Artificial Intelligence – Prof.Dr.Wolfgang Ertel
 Introduction

# What is Intelligence?
✓ Behave as intelligently as human (one school of thought –
accept humans are intelligent).
✓ behave in best manner or most rational behavior (other school
of thought – however, as human would not behave
intelligently all the time).

# behavior
✓ thinking behavior – reasoning properly to come of with best
solution to the problem.
✓ acting behavior – acting in best possible manner for the best
solution.
 Introduction
# Intelligence:
Roberto Colom, 2010 – Human Intelligence and Brain Networks:
Intelligence can be defined as a general mental ability for
reasoning, problem solving, and learning. Because of its general
nature, intelligence integrates cognitive functions such as
perception, attention, memory, language, or planning.
On the basis of this definition, intelligence can be reliably
measured by standardized tests with obtained scores predicting
several broad social outcomes such as educational achievement,
job performance, health, and longevity.

# Artificial Intelligence(AI) [cognition = knowledge]:

✓ AI is defined as the science of making computers do things that
requires intelligence when the same things are done by human.
✓ Machines try to mimic human intelligence/Machines mimic
human cognitive functions such as learning and problems solving.
 Introduction...

# Artificial Intelligence:...
In a broad sense, AI can be referred as the simulation of human
intelligence in machines that are programmed to think like humans
and mimic their actions.
• AI refers to the simulation or approximation of human
intelligence in machines.
• The goals of artificial intelligence include computer-enhanced
learning, reasoning, and perception.
• Weak AI tends to be simple and single-task oriented, while
strong AI carries on tasks that are more complex and
human-like.
• AI is being used today across different industries from finance
to healthcare.
 Introduction...

Strong Artificial Intelligence

Strong artificial intelligence (AI), also known as artificial general
intelligence (AGI) or general AI, is a theoretical form of AI used to
describe a certain mindset of AI development.

If researchers are able to develop Strong AI, the machine would

require an intelligence equivalent to humans; it would have a
self-aware consciousness that has the ability to solve problems,
learn, and plan for the future.

From the definition, Strong AI is expectation; Strong AI will be

achieved when machines have the cognitive functionalities/abilities
that humans have without any human intervention.
 Introduction...

Weak Artificial Intelligence

Weak AI, also known as Narrow AI, focuses on performing a
specific task, such as answering questions based on user input or
playing chess. It can perform one type of task, but not both,
whereas Strong AI can perform a variety of functions, eventually
teaching itself to solve for new problems.

Weak AI relies on human interference to define the parameters of

its learning algorithms and to provide the relevant training data to
ensure accuracy.

All the AI systems in existence today are Weak AI; Self-driving cars
and virtual assistants, like Siri, are examples of Weak AI.
 Introduction...

John McCarthy – one of the AI Pioneers, in 1955:

“Artificial Intelligence is the science and engineering of
making machine intelligent.”
Roughly it can be restated as: “The goal of AI is to develop
machines that behaves as though they were intelligent.”

Definition of AI found in Encyclopedia Britannica:

“AI is the ability of digital computers or computer controlled
robot to solve problems that are commonly associated with
the higher intellectual processing capabilities of human.”

Elaine Rich in 1985 - Artificial Intelligence and Humanities:

“Artificial Intelligence is the study of how to make computers
do things at which, at the moment, people are better.”
 Introduction...

Unsurprisingly, Some of the daily activities with AI tie-up:

✓ When using Google search engine [RankBrain-2015,

BERT-2019, multitask unified model (MUM)-2021] that
recommends words or combination of words.
✓ While using Netflix or Youtube, recommendation systems (an
AI application) recommend movies or videos of our taste.
✓ Siri, Apple; Alexa, Amazon; and Google Assistant/Ok Google,
Google are another AI-powered virtual assistants to process
voice inquiries, which uses natural language processing.
✓ Getting notifications in Facebook, Twitter, and Instagram for
the timeline post by a user for which we mostly view or
comment.
 Introduction...

# Approaches to AI
Thought/Reasoning

Systems that think like Systems that think

human rationally
Cognitive science Laws of thought/logic

Human like Ideal Performance

Performance Rationality

Turing Test Rational agent

Systems that act like Systems that act
human rationally

Behaviour
 Introduction...
Systems that think humanly Systems that think rationally
“The exciting new effort to make “The study of mental faculties
computer think... machines with through the use of computational
minds, in the full and literal sense” models”– Charniak and McDer-
- Haugeland, 1985. mott, 1985. (Reason, Memory,
“Automation of Activities that we as- Imagination, Perception)
sociate with human thinking, prob- “The study of computations that
lem solving, learning...”– Bellman, make it possible to perceive, reason,
1978. and act”– Wiston, 1992.
System that acting humanly Systems that acting rationally
“The art of creating machines “Computational intelligence is the
that perform functions that re- study of design of intelligent
quire intelligence when performed agent” – Poole et. al., 1998.
by people.”– Kurzweil, 1990 “AI... is concerned with intelli-
“The study of how to make computer
gent behavior in artifact”– Nils-
do things at which at the moment
people are better.”– Rich & Night, son, 1998
1991.
 Artificial Intelligence Approaches

Thinking Like Human - cognitive modeling approach

- It needs understanding the

phenomena of human
thinking activities such as
decision making, problem
solving and learning etc.
- the phenomena is to be
implemented to the AI Fig. 1 Human brain vs Computing
system; therefore, needs System
overlapping of human brain
and AI computing system.

¬ Figure displayed shows non-overlapping but approaching human

brain and computing system.
 Artificial Intelligence Approaches

Acting Like Human - Turing Test approach

- If the reply to human

interrogator makes the
interrogator confused
whether reply from human
or AI system, the AI system
action is said to be
equivalent to human action.
Fig. 2 Turing Test Approach
- for this AI System needs
almost all knowledge as by
human.
 Turing Test approach...

Factors required to pass the Turing Test:

✓ Natural Language Processing(NLP): it enables AI system
communicate easily and successfully.

✓ Knowledge Representation: to store information in the

form easy to reason.

✓ Automated Reasoning: AI system should use the stored

information to answer questions and to make inference.

✓ Machine Learning: AI should adapt to new environment or

circumstances and detect extrapolate patterns.
 Artificial Intelligence Approaches
Thinking Rationally - The Laws of Thought approach
How can machine think Rationally?
If a machine is able to develop right logic, then machine can be
able to think rationally (does right thing for what it knows).
- Logicians in 19th century Socrates is a man.
developed different logics All men are mortal.
that can in principle solve Hence, Socrates is mortal.
any problem described in
logical notations.
- Hard to model informal
knowledge, e.g., probability,
into standard logic.
- Solve a problem in principle
vs. in practice.
Fig. 3 Laws of Thought Approach
 Artificial Intelligence Approaches

Acting Rationally - Rational Agent approach

- An agent (from Latin word agere, to do) is something (e.g.,
program, robot and even human) that acts.
- A rational agent is one that acts to achieve the best outcome.
- Our course concentrate on this category of AI, which includes
studying general principles of rational agents and components
for constructing them.
* It builds on mathematical rationality, not human thinking or
acting.
* It extends the “laws of thinking” because reasoning logically is
part of being a rational agent.
E.g., retreating from hot stove is a quick reflex action.
Brief History of Artificial Intelligence
Brief History of Artificial Intelligence...

Fig. 4 How AI evolved to today’s industrial applications

(Ref. Queensland Brain Institute - UoQ)
 Brief History of Artificial Intelligence

Fig. 5 How AI evolved to today’s industrial applications

 Brief History of Artificial Intelligence

Fig. 6 How Neural Network evolved to today’s industrial

applications
 Artificial Intelligence techniques

1 Symbolic AI (¬ which has roots in logic)

2 Connectionist AI (¬ which has roots in Neuroscience)

Symbolic AI
- more commonly known as rule-based AI, or good
old-fashioned AI (GOFAI) or classic AI.
- Earlier AI development and research was based on Symbolic
AI inserting human behavior and knowledge in the form of
computer codes and processed by expert system.
- Symbolic AI system has a combination of characters signifying
real-world concepts or entities through a series of symbols.
- These symbols can easily be arranged through networks and
lists or arranged hierarchically telling AI algorithms how each
symbol is related to each other in totality.
 Artificial Intelligence techniques

Connectionist AI/Non-Symbolic AI
- it is also known as neural AI and more popular these days.
- Neural AI is based on how a human mind functions and its
neural interconnections; that is neural network.
- Neural Network is made up of hundreds of single processing
units called neurons.
- They are a layered format forming connections with weights
(parameters) which are adjustable.
Importance and Applications of Artificial Intelligence
Scope of Artificial Intelligence

Artificial Intelligence scope can be studied as:

Fig. 7 Scope of Artificial Intelligence

Importance and Applications of Artificial Intelligence...

One perspective of classifying AI applications:

1 Learning
2 Reasoning
3 Problem solving
4 Language understanding/Linguistic Intelligence
5 Perception

Another perspective of classifying AI applications:

1 Mundane Task (¬ perception-vision, speech, NLP)
2 Formal Task (¬ games, maths, geometry, logic, integral)
3 Expert Task (¬ financial analysis, medical diagnosis)
 Applications of of Artificial Intelligence...

Components of Artificial Intelligence

Fig. 8 Components of Intelligence

 Applications of Artificial Intelligence...

Learning
- it means to acquire new
knowledge or things.
- it provides system the
learning ability to
automatically learn from the
data (experience) and
improve performance
without being explicitly Fig. 9 Learning(Human & Machine)
programmed.

✓ Learning that involves generalization performs better in

situations not previously encountered.
 Applications of Artificial Intelligence...

Reasoning
- Reasoning means to infer
facts from given facts.
- Reasoning is a form of
inference and logical
techniques that is expressed
through words (qualitative)
and other symbols
(quantitative reasoning). Fig. 10 Reasoning in Intelligence

✓ Reasoning in artificial intelligence helps machines think

rationally and perform functions like humans.
✓ Probabilistic Reasoning in Artificial Intelligence allows machines
to deal with and represent uncertain knowledge and information.
 Applications of Artificial Intelligence...

Linguistic Intelligence
- Language understanding
means to understand natural
Language meaning.
- It is one’s ability to use,
comprehend, speak, and
write the verbal and written
language.
- in AI, it means understand
user speech, written
materials and respond
Fig. 11 Language Understanding
accordingly.
✓ In broad sense, Language Understanding in artificial intelligence
helps machines for ⃝1 machine translation, ⃝
2 sentiment analysis,
⃝
3 information extraction, ⃝
4 question answering.
 Applications of Artificial Intelligence...

Problem Solving
- Problem-solving means to
move towards the goal.
- Psychologically, it refers to a
state where we wish to
reach to a definite goal from
a present state or condition.”
- in AI, problem-solving
encompasses a number of
techniques such as search
algorithms, heuristics to Fig. 12 Problem Solving
solve a problem.

✓ It is the process in which one perceives and tries to arrive at a

desired solution from a present situation by taking some path.
 Applications of Artificial Intelligence...

Perception
- It is the process of acquiring,
interpreting, selecting, and
organizing sensory information.
- Perception presumes sensing.
- In humans, perception is aided by
sensory organs.
Fig. 13 Perception in Intelligence

✓ In the domain of AI, perception mechanism puts the data

acquired by the sensors together in a meaningful manner.
 Applications of Artificial Intelligence...
Applications of Artificial Intelligence

Fig. 14 Applications of Artificial Intelligence

 Applications of Artificial Intelligence...
AI in Astronomy
- Artificial Intelligence can be very useful to solve complex
universe (astronomical) problems.
- AI technology can be helpful for understanding the universe
such as how it works, origin, etc.
AI in Healthcare
- For the last decade, AI is becoming more advantageous for
the healthcare industry and predicted to have significant
impact to the industry. ¬ diabetic retinopathy diagnosis,
cancer classification, covid diagnosis.
- Healthcare Industries are applying AI to make a better and
faster diagnosis than humans.
- AI can help doctors with diagnoses and can inform when
patients are worsening so that medical help can reach to the
patient before hospitalization.
 Applications of Artificial Intelligence...

AI in Gaming
- AI can be used for gaming purpose.
- The AI machines can play strategic games like chess, where
the machine needs to think of a large number of possible
places.

AI in Finance
- AI and finance industries are the best matches for each other.
- The finance industry is implementing automation, chatbot,
adaptive intelligence, algorithm trading, and machine learning
into financial processes.
 Applications of Artificial Intelligence...

AI in Data Security
- The security of data is crucial for every company and
cyber-attacks are growing very rapidly in the digital domain.
- AI can be used to make your data more safe and secure.
- Some examples such as AEG bot, AI2 Platform, are used to
determine software bug and cyber-attacks in a better way.

AI in Social Media
- Social Media sites such as Facebook, Twitter, and Snapchat
contain billions of user profiles, which need to be stored and
managed in a very efficient way.
- AI can organize and manage massive amounts of data.
- AI can analyze lots of data to identify the latest trends,
hashtag, and requirement of different users.
 Applications of Artificial Intelligence...
AI in Travel & Transport
- AI is becoming highly demanding for travel industries.
- AI is capable of doing various travel related works such as
from making travel arrangement to suggest the hotels, flights,
and best routes to the customers.
- Travel industries are using AI-powered chatbots which can
make human-like interaction with customers for better and
fast response.

AI in Automotive Industry
- Some Automotive industries are using AI to provide virtual
assistant to their user for better performance.
- Tesla has introduced TeslaBot, an intelligent virtual assistant.
- Various Industries are currently working for developing
self-driven cars making journey more safe and secure.
 Applications of Artificial Intelligence...

AI in Robotics
- Artificial Intelligence has a remarkable role in Robotics.
Usually, general robots are programmed such that they can
perform some repetitive task, but with the help of AI, we can
create intelligent robots which can perform tasks with their
own experiences without pre-programmed.
- Humanoid Robots are best examples for AI in robotics,
recently the intelligent Humanoid robot, Erica and Sophia,
has been developed which can talk and behave like humans.

AI in Entertainment
- We are currently using some AI based applications in our daily
life with some entertainment services: Netflix or Amazon.
- With the help of ML/AI algorithms, these services show the
recommendations for programs/shows or videos.
 Applications of Artificial Intelligence...

AI in Agriculture
- Agriculture is an area which requires various resources, labor,
money, and time for best result.
- Now a day’s agriculture is becoming digital, and AI is
emerging in this field.
- Agriculture is applying AI as agriculture robotics, solid and
crop monitoring, predictive analysis.
- Therefore, AI in agriculture can be very helpful for farmers.

AI in E-commerce
- AI is providing a competitive edge to the e-commerce
industry, and it is becoming more demanding in the
e-commerce business.
- AI is helping shoppers to discover associated products with
recommended size, color, or even brand.
 Applications of Artificial Intelligence...

AI in education
- AI can automate grading so that the tutor can have more
time to teach.
- AI chatbot can communicate with students as a teaching
assistant.
- AI in the future can be work as a personal virtual tutor for
students, which will be accessible easily at any time and from
any place.
 Knowledge and Learning
# Data:
- are streams of raw facts obtained through acquisition system.
- data represents process or phenomena (stochastic process).
- can be arranged into the form understandable and usable.

# Information:
- is processed and organized data .
- is structured to give context and meaning.
- in other words, data shaped into form meaningful
(understandable, interpretable) and usable to human.

# Knowledge(Psychological result of perception; learning and reasoning):

- is what we know, and is unique to individual.
- is accumulation of past experience and insights that could
reshape the interpretation and assign meaning to information.
 Knowledge and Learning...

Fig. 15 Transformation of data to knowledge through information

 Knowledge and Learning...

# Cognitive Process:
Before defining Learning, lets talk about “cognitive process:”
¬ Physically, cognitive processes are series of chemical and
electrical signals stimulated in brain that allows to comprehend
your environment and gain knowledge.
Neurons release chemicals that create electrical signals in nearby
neurons, creating mass of signals that are then translated into
conscious and unconscious thoughts.
Denise Krch, 2011: “Cognitive processing is (a general term to
describe) a series of cognitive operations carried out in the creation
and manipulation of mental representations of information.”
Cognitive Process includes: ¬ attention, perception, reasoning,
emotion, learning, synthesizing, manipulation of stored
information, memory storage, retrieval.
 Knowledge and Learning...

# Learning:
• it is cognitive process of acquiring skill or knowledge.
• it makes useful change in our mind with mental representation
of information(mental state).
• it constructs or modifies the representation of what is being
experienced.
Herbert Simon, 1983: Learning is the phenomenon of knowledge
acquisition in the absence of explicit programming.
Learning involves three factors: Changes, Generalization, and
Improvement.
 Knowledge and Learning...

Changes:
Learning changes learner. In machine learning, problem is
determining the nature of following changes (random process) and
represent them in the best form (✓model the distribution).

Generalization:
AI/learning, in fact, is Generalization where performance must
improve not only on the same task but on similar task.

Improvement:
Learning leads to the improvements. It must address the possible
degradation of performance for some changes. Generally, Learning
should improve performance.
 Intelligent Agent

# Agent
An agent is something that can be viewed as perceiving its
environment through sensors and acting up on that environment
through effectors/actuators.

In electrical and computer term, agents are the combination of

architecture and programs. The agent program is a function that
implement the agent mapping true percepts (percept sequence) to
action through actuators. The program runs devices with sensors
and actuators which is called architecture.

Mathematically, the task or behavior of agent is to described by

agent function which is implemented as agent-program.
 Intelligent Agent...

Human Agent
Eyes, ears and other organs for sensors; hands legs, mouth and
another body parts for actuators.

Robotic Agent
Cameras and infrared for sensor; various motor for the actuators.

Software Agent
It receives keystrokes, file contents, and network packets as sensory
inputs; (acts on the environment by) displaying on the screen,
writing files, and sending network packets are action by
effectors/actuators.
 Intelligent Agent...

Fig. 16 Agents interacting with environment

• The term“percept” refers to as the agent’s perceptual input at

any given instant – the location and state of the environment.
• An agent’s percept sequence is the complete history of
everything the agent has ever perceived.
• The agent function maps from percept histories to actions.
 Intelligent Agent...

# Intelligent Agent
An agent perceives its environment(E) via sensors(S) and acts
rationally upon environment with its actuators/effectors(A) to
maximize its performance measure(P) based on the evidence by
percepts sequence and whatever knowledge built-in the agent has
(can be called rational agent). ¬ PEAS Representation.
In terms of computer science, an intelligent agent is a computer
software system whose main characteristics are situatedness,
autonomy, adaptivity, and sociability.

Situatedness
The agent receives some form of sensory input from its
environment, and it performs some action that changes its
environment in some way. Examples of environments: the physical
world and the Internet.
 Intelligent Agent...

Autonomy
The agent can act without direct intervention by humans or other
agents; it has control over its own actions and internal state.

Adaptivity
The agent is capable of
✓ (1) reacting flexibly to changes in its environment;
✓ (2) taking goal-directed initiative (i.e., is pro-active), when
appropriate; and
✓(3) learning from its own experience, its environment, and
interactions with others.

Sociability
The agent is capable of interacting in a peer-to-peer manner with
other agents or humans.
 Intelligent Agent...

There are five basic types of agents:

1 Table-driven agents

2 Simple reflex agents

3 Model-based reflex agents

4 Goal-based agents

5 Utility based agents

 Intelligent Agent...

Table-driven agents:
• it uses a percept
sequences/action pairs in
memory to find the next
action to be performed.
• they are implemented by a
lookup table.
• it takes long time to build
lookup table.
• not adaptable to the change
environment, requires Fig. 17 Table-driven agent
change in lookup table.
 Intelligent Agent...

Simple Reflex agents:

• agents are based on

condition-action (if-then)
rules and immediate percept.
• they are stateless devices
which do not have a
memory of the past state.
• environment should be fully
observable.
Fig. 18 Simple-Reflex agent
 Intelligent Agent...

Model based Reflex agents:

• model based means
knowledge based agent.
• agents with memory have its
internal state .
• next action will be based on
present and past experience
or state(percept history).
• it is partially observable.
Fig. 19 Model based Reflex agents
 Intelligent Agent...

Goal based agents:

• expansion of model based
agent (with percept history).
• agents have state
information and goal
information (supervised).
• goal information describes
the desirable situation.
• agents of this kind has to
consider future events.
• searching and planning to Fig. 20 Goal based agents
reach goal state.
 Intelligent Agent...

Utility-based agents:

• decisions are based on

classic axiomatic utility
theory.
• deals with happy and
unhappy states.
• utility function measures if
the agent is in happy state.
• it is partially observable.

Fig. 21 Utility based agents

 Module Assignment – As You Go

Module#1 Assignment is available at MS-Team.

Submission Deadline: 27th January 2023 (Before 3:00 PM)