UNESCO Chair

Artificial Intelligence:
Introduction and Applications
Salah Obayya, PhD, DSc
Fellow IEEE, OPTICA (formerly OSA), SPIE, APS, AAS,
ACES, HEA, IEE, IOP
Professor & Founding Director of Centre for Photonics and Smart Materials (CPSM)
Vice Chairman, Zewail City (formerly)
Zewail City of Science and Technology, Egypt,
UNESCO-Chair Professor, Paris, France
&
Honorary Chair Professor of Photonics, University of Nottingham, UK
1
Salah Obayya
UNESCO Chair

Outline
• AI: Definition, History, Types, and Domains
• Machine Learning
• Deep Learning “Neural Networks”
• Natural Language Processing (NLP)
• Computer Vision (CV)
• Robotics
• Ethics and Social Implications
• Current and Future Applications
• Conclusion

Salah Obayya
UNESCO Chair

What is AI?
• AI stands for Artificial Intelligence, which refers to the ability of
machines to perform tasks that typically require human intelligence.

3
Salah Obayya
UNESCO Chair

What is AI?
 AI systems are designed to simulate human intelligence and behavior,
and they can be trained to recognize patterns, make predictions, and
automate tasks based on input data.

4
Salah Obayya
UNESCO Chair

History of AI
 The history of AI dates back to ancient times, when philosophers and
inventors began exploring the concept of artificial beings and intelligent
machines.

 Yan Shi was a Chinese

inventor who lived during the
5th century BC. He designed
a humanoid automaton that
could perform various tasks
such as walking, singing, and
dancing.

5
Salah Obayya
UNESCO Chair

History of AI
 The history of AI dates back to ancient times, when philosophers and
inventors began exploring the concept of artificial beings and intelligent
machines.

 Al-Jazari was a Muslim inventor who

lived during the 12th century AD.
He designed a number of automata,
including a humanoid musician, a
mechanical peacock, and a water clock
that featured humanoid figures that
moved and played music.

6
Salah Obayya
UNESCO Chair

History of AI
 The history of AI dates back to ancient times, when philosophers and
inventors began exploring the concept of artificial beings and intelligent
machines.

 Leonardo da Vinci was an Italian artist

and inventor who lived during the 15th
and 16th centuries. He designed a
number of robotic devices, including a
mechanical knight that could sit up,
wave its arms and move its head.

7
Salah Obayya
UNESCO Chair

History of AI
 In the 1940s, The term "artificial intelligence" is coined by John
McCarthy, Marvin Minsky, and other researchers at Dartmouth College.
The field of AI is officially established with the goal of creating machines
that can think and learn like humans.

 In the 1960s, The first AI systems, that were developed, could understand
and respond to natural language. The ELIZA program, developed by
Joseph Weizenbaum, is one of the first chatbots that can simulate a
conversation with a human user.

 In the 2000s, Deep learning techniques were developed, which are based
on neural networks with many layers.

8
Salah Obayya
UNESCO Chair

History of AI
 Today, AI is used in a wide range of applications, including self-driving
cars, virtual assistants, and medical diagnosis. As AI technology continues
to advance, it has the potential to revolutionize many industries and
change the way we live and work.

9
Salah Obayya
UNESCO Chair

AI: Classification
AI

Artificial Artificial Artificial

Narrow General Super
Intelligence Intelligence Intelligence

 Artificial Narrow Intelligence (ANI) refers to AI systems that are designed to

perform a specific task or set of tasks within a narrow domain.

 These systems are highly specialized and can only perform the tasks they are
programmed to do.

10
Salah Obayya
UNESCO Chair

AI: Classification
AI

Artificial Artificial Artificial

Narrow General Super
Intelligence Intelligence Intelligence

 Artificial General Intelligence (AGI) refers to AI systems that have the ability to
understand or learn any intellectual task that a human can.
 These systems would be capable of reasoning, problem-solving, and learning in a
way that is similar to humans.
 However, AGI is still a theoretical concept and has not yet been achieved.
11
Salah Obayya
UNESCO Chair

AI: Classification
AI

Artificial Artificial Artificial

Narrow General Super
Intelligence Intelligence Intelligence

 Artificial Super Intelligence (ASI) refers to AI systems that are more intelligent than
humans in every way.
 These systems would be capable of solving problems and creating innovations that
are beyond human comprehension.
 There is much debate among experts about whether it is even possible to achieve.
12
Salah Obayya
UNESCO Chair

AI Domains
 Artificial Intelligence (AI) is a vast field that encompasses a range of
domains, including Machine Learning (ML), Deep Learning “Neural
Networks”, Natural Language Processing (NLP), Computer Vision (CV),
and Robotics.

13
Salah Obayya
UNESCO Chair

Machine Learning
 Machine learning is a subset of AI that involves creating algorithms and
models that can learn from data and improve their performance over
time.

 ML algorithms can identify patterns and relationships in large datasets,

and use these patterns to make predictions or take actions without being
explicitly programmed to do so.

 Machine learning is used in a variety of applications, including image

recognition, natural language processing, recommendation systems, and
predictive analytics.

14
Salah Obayya
UNESCO Chair

Machine Learning
 There are three types of Machine Learning: Supervised Learning,
Unsupervised Learning, and Reinforcement Learning (RL).

15
Salah Obayya
UNESCO Chair

Supervised Machine Learning

 Supervised Machine Learning is a type of machine learning in which the
algorithm is trained on a labeled dataset, where the input data is paired
with corresponding output data.

 The goal of supervised learning is to learn a mapping function from the

input data to the output data, so that the algorithm can make accurate
predictions on new, unseen input data.

 The labeled dataset used in supervised learning typically consists of two

parts: the input data (also known as features) and the corresponding
output data (also known as labels or targets).

16
Salah Obayya
UNESCO Chair

Supervised Machine Learning

 The algorithm is trained by iteratively adjusting its parameters to
minimize the difference between its predicted output and the actual
output in the labeled dataset.

 Supervised learning can be further classified into two main types:

 Regression: In regression, the output data is continuous, and the goal is to learn
a function that can predict a numeric value for new input data. Examples of
regression problems include predicting housing prices or stock prices.

 Classification: In classification, the output data is categorical, and the goal is to

learn a function that can predict a class label for new input data. Examples of
classification problems include spam detection, image classification, and
sentiment analysis.

17
Salah Obayya
UNESCO Chair

Unsupervised Machine Learning

 Unsupervised Machine Learning is a type of machine learning in which the
algorithm is trained on an unlabeled dataset, where there is no corresponding
output data.

 The goal of unsupervised learning is to identify patterns and relationships in the

input data without any guidance, and to create a compact representation of the
input data.

 Unsupervised learning is widely used in a variety of applications, including data

mining, image and speech processing, and recommendation systems.

 However, it can be difficult to evaluate the performance of unsupervised

learning algorithms, as there is no clear objective function to optimize.

18
Salah Obayya
UNESCO Chair

Unsupervised Machine Learning

 Supervised learning can be further classified into three main types:

 Clustering: In clustering, the algorithm groups similar data points together into
clusters based on some similarity metric. Examples of clustering techniques
include k-means clustering and hierarchical clustering.

 Dimensionality reduction: In dimensionality reduction, the algorithm reduces

the number of features in the input data while preserving as much information
as possible. Examples of dimensionality reduction techniques include principal
component analysis (PCA) and t-distributed stochastic neighbor embedding (t-
SNE).

 Anomaly detection: In anomaly detection, the algorithm identifies data points

that are significantly different from the rest of the data. Anomaly detection can
be used for tasks such as fraud detection and fault diagnosis.

19
Salah Obayya
UNESCO Chair

Reinforcement Learning

 Reinforcement Learning (RL) is a type

of Machine Learning in which an
agent learns to make decisions by
interacting with an environment and
receiving feedback in the form of
rewards or penalties.

 The goal of RL is to learn a policy,

which is a mapping from states to
actions, that maximizes the expected
cumulative reward over time.

20
Salah Obayya
UNESCO Chair

Reinforcement Learning

 In RL, the agent takes an action based

on its current state, and the
environment responds with a new
state and a reward signal that reflects
the quality of the action taken.

 The agent then updates its policy

based on the observed state, action,
and reward, and repeats the process
of taking actions and receiving
feedback until it reaches a terminal
state.

21
Salah Obayya
UNESCO Chair

Reinforcement Learning

 In RL, the system doesn’t require any

input data “labeled or unlabeled”,
however, it generates its own training
data while interacting with the
environment which the agent uses to
assess the quality of its actions.

 RL can be used to solve a variety of

problems, including game playing,
robotics, and recommendation
systems.

22
Salah Obayya
UNESCO Chair

Neural Networks
 Neural Networks are a type of machine learning algorithm inspired by the
structure and function of the human brain.

 They are composed of layers of interconnected nodes, or neurons, that

process information and learn to make predictions or take actions based
on input data.

 Each neuron in a neural network receives input from other neurons,

applies an activation function to the input, and produces an output that
is passed on to other neurons in the next layer.

 Activation functions are mathematical functions used in Neural Networks

to introduce non-linearity into the output of each neuron.
23
Salah Obayya
UNESCO Chair

Neural Networks
 The weights and biases of the connections between neurons are adjusted
during training to optimize the network's performance on a specific task.

Single Neuron

24
Salah Obayya
UNESCO Chair

Neural Networks: FeedForward

 In a feedforward neural network, information flows in one direction from
input to output, through one or more hidden layers. These networks are
widely used for tasks such as image classification, speech recognition,
and natural language processing.

25
Salah Obayya
UNESCO Chair

Neural Networks: Convolution

 Convolutional Neural Networks (CNNs) are a specialized type of feedforward
neural network that are designed to process grid-like data, such as images. They
use convolutional layers to extract local features, such as edges and textures,
and pooling layers to down sample the feature maps.

26
Salah Obayya
UNESCO Chair

Neural Networks: Recurrent

 Recurrent Neural Networks (RNNs) are designed to process sequential data,
such as time series or natural language text. They use recurrent connections to
maintain state information across time steps and can be used for tasks such as
speech recognition, language modeling, and sentiment analysis.

27
Salah Obayya
UNESCO Chair

Neural Networks: Applications

28
Salah Obayya
UNESCO Chair

Natural Language Processing

 Natural Language Processing (NLP) is

a field of study that focuses on the
interaction between computers and
human language. It involves the use
of computational techniques to
analyze, understand, and generate
natural language text or speech.

 However, NLP still faces challenges,

such as understanding the nuances
of human language, dealing with
ambiguity, and handling the vast
amount of data involved.

29
Salah Obayya
UNESCO Chair

Natural Language Processing

 Some of the tasks that NLP can perform include:

 Text Classification: NLP can be used to classify text into different categories, such as
spam or not spam, sentiment analysis, and topic classification.

 Named Entity Recognition: NLP can be used to identify and classify named entities in
text, such as people, organizations, and locations.

 Sentiment Analysis: NLP can be used to determine the sentiment of a piece of text,
such as whether it is positive, negative, or neutral.

 Question Answering: NLP can be used to answer questions posed in natural

language, such as those used in chatbots or virtual assistants “ChatGPT”.

30
Salah Obayya
UNESCO Chair

NLP Techniques
 Some of the techniques used in NLP include:

 Tokenization: The process of splitting text into individual words or tokens.

31
Salah Obayya
UNESCO Chair

NLP Techniques
 Some of the techniques used in NLP include:

 Part-of-Speech Tagging: The process of assigning a part of speech to each token,

such as noun, verb, or adjective.

32
Salah Obayya
UNESCO Chair

NLP Techniques
 Some of the techniques used in NLP include:

 Parsing: The process of analyzing the grammatical structure of a sentence.

33
Salah Obayya
UNESCO Chair

NLP Applications

34
Salah Obayya
UNESCO Chair

Computer Vision
 Computer vision is a field of artificial intelligence and computer
science that focuses on enabling computers to interpret, analyze,
and understand visual information from the world around us.

 It involves the use of algorithms and mathematical models to

extract meaningful information from images, videos, and other
forms of visual data.

 Advancements in computer vision have been accelerated by the

availability of large datasets, powerful computers, and deep
learning techniques such as convolutional neural networks (CNNs).

35
Salah Obayya
UNESCO Chair

Computer Vision: CNN

 We already discussed the CNN in briefly before, so, we will expand
on it.

 The basic building block of a CNN is the convolutional layer, which

applies a set of filters to an input image to extract features that are
relevant for a given task.

 Each filter is a small matrix of weights that is applied to a small

patch of the input image, and the output of each filter is a feature
map that highlights the presence of certain visual patterns or
structures in the input image.

36
Salah Obayya
UNESCO Chair

Computer Vision: CNN

 Convolution can be perceived as in the following example of
convoluting a 5x5 matrix into a 3x3 matrix using a 3x3 filter:

37
Salah Obayya
UNESCO Chair

Computer Vision: CNN

 So, if we visualize images as matrices, this will be the effect of
convolution on images:

38
Salah Obayya
UNESCO Chair

Computer Vision: Applications

39
Salah Obayya
UNESCO Chair

Robotics
 Robotics is the branch of engineering and technology that deals with the design,
construction, operation, and use of robots.

 A robot is a programmable machine that is capable of carrying out complex

tasks autonomously or under remote control.

 Advancements in robotics technology have been driven by improvements in

sensors, actuators, and control systems, as well as developments in artificial
intelligence and machine learning.

 One of the key challenges in robotics is developing robots that are capable of
interacting with the physical world in a safe and efficient manner.

40
Salah Obayya
UNESCO Chair

Robotics
 This requires developing robust control algorithms that can handle uncertainty
and variability in the environment, as well as developing sensors and feedback
systems that can provide accurate and timely information about the robot's
surroundings.

 Another important area of robotics is human-robot interaction, or HRI. HRI is

concerned with developing robots that can interact with humans in a natural
and intuitive way and is a key area of research in fields such as healthcare and
education.

 This requires developing robots that can perceive and respond to human
emotions and intentions, as well as developing interfaces that are easy to use
and understand.

41
Salah Obayya
UNESCO Chair

Robotics: Industrial Robots

 Industrial robots: These are used in manufacturing and industrial settings to
perform tasks such as welding, painting, assembly, and material handling.

42
Salah Obayya
UNESCO Chair

Robotics: Service Robots

 Service robots: These are used to provide services to humans, such as in
healthcare, education, entertainment, and personal assistance.

43
Salah Obayya
UNESCO Chair

Robotics: Military Robots

 Military robots: These are used for military applications, such as
reconnaissance, surveillance, and bomb disposal.

44
Salah Obayya
UNESCO Chair

Robotics: Agricultural Robots

 Agricultural robots: These are used in farming and agriculture to perform tasks
such as harvesting, planting, and spraying.

45
Salah Obayya
UNESCO Chair

Robotics: Exploration Robots

 Exploration robots: These are used in exploration and research applications,
such as in space exploration or deep-sea exploration.

46
Salah Obayya
UNESCO Chair

Ethical and Social Implications

 There are a lot of ethical and social implications that arises from using AI.

Problems with AI

Job
Displacement Privacy
Bias of AI
and Creation Concerns

47
Salah Obayya
UNESCO Chair

Bias of AI
Type “greatest leaders of all time” in your favorite search
engine and you will probably see a list of the world’s
prominent male personalities. How many women do you
count?

48
Salah Obayya
UNESCO Chair

Bias of AI

AI-systems deliver biased results

Why ?

 AI technology is not neutral as it processes big data and prioritizes results

with the most clicks relying both on user preferences and location.

 Thus, a search engine can become an echo chamber that upholds biases
of the real world and further entrenches these prejudices and stereotypes
online.

49
Salah Obayya
UNESCO Chair

Bias of AI
How can we ensure more
equalized and accurate results?

Bias should be avoided or at the least minimized in the development of

algorithms, in the large data sets used for their learning, and in AI use
for decision-making.

 Al Data Sets should comply with

those core values:
 Protect Human Rights
 Provide a Peaceful life
 Ensure Diversity and Inclusion
 Protect the Environment
50
Salah Obayya
UNESCO Chair

Job Displacement

51
Salah Obayya
UNESCO Chair

Job Creation

52
Salah Obayya
UNESCO Chair

Jobs in the AI Era

What Should We Do ?

 Re-training and Education Programs: Governments and companies can

fund re-training programs for workers displaced by AI. This could include
training for new jobs that complement AI, or for in-demand fields that AI
has not yet impacted significantly.

 Job Sharing and Reduced Working Hours: Instead of laying off workers as
AI takes over certain tasks, companies could implement job sharing and
reduced working hours. Workers could split jobs or work fewer days,
maintaining employment but at a reduced pay.

53
Salah Obayya
UNESCO Chair

Privacy Concerns
 Here are some key privacy concerns regarding AI:

 Data Collection - As AI systems are fueled by vast amounts of data, there are
concerns about how much data is collected and how it is used. Much of this data
collection happens without people's explicit consent.

 Data Access - Once data is collected, there are questions about who has access to it
and for what purposes. There have been cases of data being misused or leaked.

 Loss of Anonymity - As AI systems get better at identifying and tracking people,

some argue this could threaten individual anonymity which has been an important
part of privacy.

 Opacity - Many AI systems are opaque "black boxes" that people cannot fully
understand or scrutinize. This makes it hard to determine if they are violating privacy.

54
Salah Obayya
UNESCO Chair

Privacy Concerns
 Here are some key privacy concerns regarding AI:

 Biased Outcomes - AI systems can reflect and even amplify the biases in the
data used to train them. This can lead to discriminatory or unfair outcomes
that threaten people's privacy.

 Lack of Control - Individuals often have little control over how AI systems
collect and use their data. This leaves people feeling powerless to protect
their own privacy.

55
Salah Obayya
UNESCO Chair

Privacy Concerns
How can we overcome these
concerns?

Individuals should be empowered with

transparency and control over their personal
data. By prioritizing privacy and adopting
strong data protection policies, we can help
ensure that AI technology is developed and
used in a way that is both effective and
privacy-respecting, ultimately leading to a
future where individuals can benefit from the
transformative power of AI without
sacrificing their fundamental right to
privacy.

56
Salah Obayya
UNESCO Chair

Privacy Concerns
How can we overcome these
concerns?

Organizations and companies that use

AI must prioritize privacy and ethical
considerations in their AI systems'
design and implementation. This
includes being transparent about data
collection and usage, ensuring data
security, regularly auditing for bias and
discrimination, and designing AI systems
that adhere to ethical principles.

57
Salah Obayya
UNESCO Chair

Privacy Concerns
How can we overcome these
concerns?

Governments should implement

regulations to ensure that AI is
developed and used in a way that
respects individual privacy and other
ethical considerations.

58
Salah Obayya
UNESCO Chair

AI Applications: Healthcare

59
Salah Obayya
UNESCO Chair

AI Applications: Finance

60
Salah Obayya
UNESCO Chair

AI Applications: Transportation

61
Salah Obayya
UNESCO Chair

AI Applications: Education

62
Salah Obayya
UNESCO Chair

AI Applications: Entertainment

63
Salah Obayya
UNESCO Chair

AI Applications: Engineering

64
Salah Obayya
UNESCO Chair

Conclusions

 Artificial intelligence (AI) is a rapidly evolving field with the potential

to revolutionize many aspects of our lives.
 Machine learning is a subset of AI that allows computers to learn
from data and improve their performance over time.
 Deep learning is a type of machine learning that uses artificial neural
networks to learn complex tasks.
 Natural language processing (NLP) is a field of AI that deals with the
interaction between computers and human language.
 Computer vision (CV) is a field of AI that deals with the ability of
computers to see and understand the world around them.
 Robotics is a field of engineering that deals with the design,
construction, operation, and application of robots.
 The ethics and social implications of AI are complex and still being
debated.

Salah Obayya
 Experience UNESCO
UNESCO Chair

Thank You

66
Salah Obayya