Sourabh Sharma, Poonam Chaudhary

Chapter 4
Machine learning and deep learning
Abstract: From the evolution of the human race, the stories that have engulfed the
curiosity of all the civilizations consist of nonliving things acting like human beings
or, in simple terms, equating humans in terms of wisdom and adapting human behav-
ior. Stories were plotted around objects like talking trees, dolls, and statues that could
talk and behave like human beings, who actually accompanied or even guided the liv-
ing entity to success, leading to a happy ending.
This is how the early concept of Artificial Intelligence (AI) came into imagination;
it was later extended by Alan Turing and, most recently, the Hollywood movie charac-
ter Iron Man introduced to the kids and youngsters through the AI bot Zarvis, who
accompanies the lead character to gain superpowers through technology.
AI has now become an integral part of the world, knowingly or unknowingly, it
has made huge advancements especially through the evolution of machine learning
(ML) and deep learning (DL) algorithms and rich computation resources. The music
that you listen to on YouTube is being used to train an ML model, which further pre-
dicts music for you; social media pages suggest different people and pages according
to your likes and dislikes; self-driving cars, your shopping apps, and many other plat-
forms are using AI to predict consumer behavior, plan efficiently, and thus, increase
their revenue.
AI is a fast moving technology with impacts and implications on both individuals
and society. With a gentle introduction to history and some important algorithms
used in AI, from ML to DL, we will be covering the major aspects of AI. The chapter
takes a balanced look at AI, its impact on the existing world, as well as its growing
potential in future.

4.1 Introduction
Artificial means “man-made” and Intelligence means “thinking power,” so we can say
Artificial Intelligence (AI) means “a man-made thinking power.” It denotes intelligence
demonstrated by machines, as natural intelligence is to humans or animals. AI exists
when a machine can perform human tasks like learning, solving problems, reasoning,

Sourabh Sharma, Department of CSE, The NorthCap University, Haryana, India,

e-mail: sourabh.714sharma@gmail.com
Poonam Chaudhary, Department of CSE, The NorthCap University, Haryana, India,
e-mail: poonamchaudhary@ncuindia.edu

https://doi.org/10.1515/9783110791402-004
72 Sourabh Sharma, Poonam Chaudhary

and many more. In the emerging era of technologies, AI is one of the fastest-growing
fields, transforming how we work, live, and play. It is all set to create a new revolution
in the world by making intelligent machines. AI is all around; you can see it in almost
each and every domain. It has become one of the most fascinating fields in Computer
science that gets machines to work as a human. With AI, you do not need to prepro-
gram a machine to do work; instead, you create a machine with programmed algo-
rithms that can work with its own intelligence and that is the beauty of it.

4.1.1 History of AI

AI is not a newly coined term but is much older than you could have imagined. The
given timeline demonstrates the evolution of AI with time. The history of AI has had
cycles of success, misplaced optimism, and resulting cutbacks in enthusiasm and
funding [1]. There have also been cycles of introducing new creative approaches and
systematically refining new ones as shown in Figure 4.1.

Artificial Intelligence’s Inception:

– 1943: The first artificial neurons model was proposed by Warren McCulloch and
Walter Pits in which each neuron is characterized as being “on” or “off,” like a
switch, in response to stimulation by a sufficient number of neighboring neurons.
– 1950: Alan Turing proposed the Turing Test to check the machine’s ability to ex-
hibit intelligent behavior comparable to human intelligence and articulated a
complete vision of AI in his article “Computing Machinery and Intelligence.”
– 1951: First neural network computer named SNARC was built by Marvin Minsky
and Dean Edmonds.

The Birth of the Artificial Intelligence:

– 1955: The first artificial intelligence program named Logic Theorist was created
by Allen Newell and Herbert A. Simon that proved 38 out of 52 mathematical
theorems.
– 1956: Artificial Intelligence was first adopted at the Dartmouth conference by
American computer scientist John McCarthy and was made a separate field.

The Growth of Artificial Intelligence:

– 1959: Geometry Theorem Prover, made by Herbert Gelenter was able to prove
many mathematical theorems.
– 1966: The first chatbot named ELIZA was created by Joseph Weizenbaum.
– 1972: The first intelligent humanoid robot named WABOT-1 was built in Japan.
 Chapter 4 Machine learning and deep learning 73

Artificial Intelligence's Inception

1943 First Aritificial Neuron Model

Turing Test 1950

1951 SNARC

The Birth of Artificial Intelligence

1955 Logistic Theorist

Artificial Intelligence
coined 1956

The Growth of Artificial Intelligence

1959 Geometry Theorem Prover

First Chatbot: ELIZA 1966

First Intelligence Humanoid
1972
Robot: WABOT-1
Reality Check 1970s

A boom of AI

1980–1987 Expert System

Fifth Generation Project 1981

1980s AI Industry boomed

The period of AI Winter 1987–1993

The Rise of Intelligent agents

First computer to beat a world chess
1997 champion: IBM Deep Blue
Roomba 2002

2006 AI Stepped into Business World

Emergence of AI

2011 IBM's Watson wins a Quiz show

Google now 2012

2014 Eugene Goostman

IBM's Project Debater 2018

Figure 4.1: Timeline of AI evolution.

74 Sourabh Sharma, Poonam Chaudhary

Reality Check:
– 1970s: During this time computer scientists dealt with a huge shortage of funding
for AI research from the government, as the machines were not giving desirable
and expected results and the money spent was huge.

A boom of AI:
– 1980–1987: Expert System brought back AI. These systems were programmed to
imitate the decision-making ability of a human expert.
– 1981: Fifth Generation project was announced, a 10-year plan to build intelligent
computers running for a prolonged period.
– The AI industry overall boomed from a few millions dollars to billions of dollars.

The period of AI winter:

– 1987–1993: Funding and investment again stopped for the AI research because of
high cost and not getting efficient results, and many companies suffered as they
failed to deliver on extravagant promises.

The rise of Intelligent agents:

– 1997: IBM Deep Blue becomes the first computer to beat the world chess cham-
pion Gary Kasparov.
– 2002: AI made its entry to homes in the form of Roomba, a vacuum cleaner.
– 2006: AI stepped into the business world when companies like Facebook, Netflix,
and Twitter started using it.

Emergence of AI:
– 2011: IBM’s Watson won a quiz show named Jeopardy, where it solved the com-
plex questions as well as riddles. It proved that natural language could be under-
stood and tricky questions can be solved easily by a machine.
– 2012: Google Now, an android app was launched by Google to provide informa-
tion to the user as a prediction.
– 2014: Chatbot Eugene Goostman won a competition of the Turing test.
– 2018: IBM’s Project Debater debated with two master debaters and performed
unbelievably well.

AI has advanced more rapidly in the past decade, with the coming of DL and Big data
and Data science. The subfields of AI have become more integrated and AI has found
common ground with other disciplines. Companies are working with AI to create
amazing devices.
 Chapter 4 Machine learning and deep learning 75

4.1.2 Advantages and disadvantages of AI

The advantages of AI [2] are countless, but to name a few:

1. Less error, high accuracy: Humans are known to be statues of mistakes, but this
can be avoided in machines if they are properly programmed. As the decisions
are made on the previously gathered information applying a certain set of algo-
rithms, errors are reduced and chances of reaching higher accuracy increase.
2. Useful in risky areas: This is the major advantage the AI-developed robots are
providing. We are able to reach Mars, even the deepest parts of oceans are discov-
ered, and robots are helping humans reach where they are not capable of or are
very risky zones.
3. Daily applications: Siri, Cortana, and OK google are used very frequently and
carry out many of the command tasks, be it making a call, sending a mail, or
many more.
4. Digital assistants: You must have seen chatbots, the best example you see around
on many websites. It can interact with humans and resolve their problems with-
out letting us know who you are talking to – a machine or a person. It saves the
need for a human resource for such tasks.
5. Easy repetitive tasks: You can easily automate the daily repetitive and boring
tasks for humans and keep them free for more creative tasks, as machines are
still not capable of using their creativity.
6. Availability 24 × 7: Humans, on average, can work for 6 h excluding the breaks.
They even need proper sleep and refreshment time to get their brains to function
in a healthy way, but by using AI, you can make machines work 24 × 7 tirelessly,
without any breaks. This has been a plus point in helpline centers that address
people’s queries, 24 × 7.

With pros come the cons, the same goes with AI. Despite making our lives easy, it is
even making us get second thoughts about its use due to the following reasons:
1. Emotionless: No doubt machines can work very efficiently, but they cannot be a
full replacement for humans. They cannot work when it comes to team manage-
ment tasks, brainstorming sessions, decision-making, and many more.
2. High cost: To be in touch with the AI updates, you have to get the hardware and
software updated from time to time for meeting the requirements; this increases
the maintenance and repair costs too.
3. Making humans dependent: AI is making humans so much dependent on them by
automating even the simpler tasks, thereby making them lazy and deteriorating
their health.
4. Lack of mental ability: Machines can only perform the work they are programmed
to do; any invalid input can lead to its failure or give us undesired outputs.
76 Sourabh Sharma, Poonam Chaudhary

5. Unemployment: Every organization is looking to automate its work, so that its

labor cost is reduced and AI robotics are helping it automate the majority of the
work that does not involve much of the mind.

Every new technology will bring both advantages and disadvantages with itself, but
humans have to take care of them and make sure to exploit only its positive side to
create a better place to live in.

4.1.3 Symbolic AI and computational AI

Symbolic AI, also called Classical AI or good old-fashioned AI, is the branch of AI that
approaches to train your machines just the way our human brain learns [3, 12–13]. It
represents problems using symbols and then uses the logic to search for the solutions.
This practice had shown impressive results in the early decades of AI research, but
has fallen wayside, today, with the uprising of neural networks that involves self
learning rules established from the correlation between the inputs and outputs.

4.1.3.1 Working

Symbolic AI takes strings of characters that represent real world entities and then
these strings are stored manually or incrementally in any appropriate data structure
[14, 15]. They are made available to the interfacing human beings or machines as and
when requested and used to make intelligent decisions and conclusions, based on the
memorized facts and rules put together by propositional logic or first-order predicate
calculus techniques. This does not require any training or massive amounts of data or
guesswork. The best part is that it can explain its decisions by showing which parts
were evaluated as true or false, unlike the black box of machine learning.

4.1.3.2 No longer popular

Symbolic AI is useful for settings where the rules are simple and straightforward, and
users can easily obtain input and transform them into symbols but fail to deal with the
messy data that exists in the real world. Thus, this technology has had minimal success,
and neural networks that imitate human intelligence have mostly taken over the field [4].
Though several efforts are being made to combine neural networks and symbolic
AI, many leading scientists still believe that symbolic AI will continue to remain a
very important part of AI. A hybrid AI system Neuro-Symbolic Concept Learner is de-
veloped by MIT_IBM Watson AI Lab that uses both rule-based programs and neural
networks to solve visual question-answering problems. This model can learn new
 Chapter 4 Machine learning and deep learning 77

tasks with fewer data, is explainable, and does not struggle with analyzing the content
of images.
Computational Intelligence (CI), also called soft computing, is the ability of a com-
puter to learn a specific task from data or experimental observations. It provides the
solution to solve complex real-world problems that cannot be solved using traditional
or mathematical modeling with the methods that are close to the human’s way of rea-
soning, i.e., it uses incomplete and inexact knowledge to produce control actions in an
adaptive way [5]. CI makes use of a combination of five main complementary techni-
ques: The fuzzy logic that makes the computer understand the natural language;, artifi-
cial neural networks that enables the system to learn experiential data by operating like
the biological one; Evolutionary computing that is based on the process of natural selec-
tion, learning theory, and probabilistic methods, which helps in dealing with uncer-
tainty and imprecision; swarm intelligence and artificial immune systems, which are
biologically inspired algorithms that can be seen as a part of evolutionary computation,
data mining, natural language processing, image processing, and AI. CI and AI scout for
similar goals to reach general intelligence, i.e., the intelligence of a machine that could
perform any intellectual task like human beings. But there is a clear distinction be-
tween them. CI is based on fuzzy logic that is much closer to the way the human brain
works, by aggregating data to partial truths – one of the main exclusive aspects of CI.
Its main application can be seen in engineering, computer science, data analysis,
and biomedicine.

4.2 Algorithms in AI
The advancements of computing infrastructure allowed the easy use of AI algorithms.
Though AI algorithms were available mathematically [6], their computation cost
was something that hindered the development of AI computation.

4.2.1 AI has found a lot of application in various fields

1. Social Media
– Identifying fraud accounts and malicious activities:
Social Media platforms have provided human society a platform to express their
thoughts and eliminate the barriers of communication [7]. They have also acted as
wheels of revolution for democratic ideas and voices. However, fake accounts and
activities with malicious intent have proved to be serious challenges; for instance,
social media platforms have become a major source of false information that re-
sult in violence and riots. AI-enabled tools are being used to detect and eradicate
fraud accounts and prevent malicious activities.
78 Sourabh Sharma, Poonam Chaudhary

– Recommendation
AI-based recommendation systems have added a shot in the arm of business intel-
ligence forming smart environments that govern the whole e-commerce system
[8]. Amazon and Flipkart have developed their AI-based recommendation systems
to better address the needs of customers and serve them better.

– Translation
Social media platforms such as Meta and Twitter have users from across the
globe. This, in turn, requires in the use of different languages; so, AI-based algo-
rithms are being used by social media platforms to translate the posts, messages,
and comments.

– Suicide prevention
Social Media has become a platform where people actively share day-to-day activ-
ities and feelings. This involves instances when people commit suicides either de-
claring it on their social media platforms or going live. Social Media giants such
as Meta are actively working on AI models to identify such instances and prevent
them by informing authorities and stakeholders on time.

2. Chat bots
Chat bots have made the web world really interactive. They not only allow the websites
to answer user specific queries on time and without the aid of human resources, saving
time, effort and money, but also enhance user experience on the website. AI-based
smart chatbots with IoT have made life easy and they are becoming helpful for people
with disabilities.

3. Finance
AI-based algorithms and tools are being used by organizations to make better deci-
sions by assessing more factors to identify and calculate financial conditions of bor-
rowers to make responsible lending. AI-based models are also being used to identify
banking frauds such as fraud credit /debit card transactions and identifying and re-
ducing risks in business.

4. Gaming
Gaming industry is also using AI models to better engage the user and provide him with
better approximations of vectors in 3D space. AI-based behavior in games generally in-
cludes responsive and intelligent activities exhibited by characters that are not players.
This results in better gaming experience and more engagements from the player
side.
 Chapter 4 Machine learning and deep learning 79

5. Automobiles
Automobile sector is one prominent sector that is leveraging AI in multiple stages,
from manufacturing to the final product. AI-based models are being used to generate
better designs and run simulations. They are also used to enhance user experience;
smart AI assistants in cars help the driver connect with the car and use the features
in an efficient way . Smart AI systems in the cars learn about the driver – his destina-
tions points, favorable car temperatures, and playlists and provide results accord-
ingly. For instance, they suggest routes with lesser traffic and taxes, etc., setting
favorable temperature and playing music that are more aligned to the driver’s pre-
ferred tastes Automobile buying and selling platforms are also leveraging AI to build
smart recommendation systems to better serve the customer.

6. Agriculture
AI has found numerous applications in the field of agriculture. There are mobile-based
applications that leverage AI algorithms to suggest best practices for farming, based on
geographical locations and season. They also provide weather forecasts. AI-based tools
are being used for crop and soil monitoring and crop yield predictions, along with
input quantity and prices.

7. Weather
Weather forecasting is crucial for every sector of the economy. It is also crucial for
mitigating the risks caused by natural calamities such as cyclones and floods. Even we
need it for our day-to-day life for scheduling our business trips, holidays, etc.
Traditional forecasting models are very heavy – there is tedious workload from
collecting the data, integrating with data from other sources, and then simulating it.
This requires higher storage and computation resources, such as supercomputers.
Still, it takes a lot of time to process it, along with IoT of power and proportionately
cost also tends to increase.
Researchers are now working on DL-based models to better learn the patterns,
which is something that AI models are highly specialized in. Results have shown that
AI-based models have shown significant improvement in terms of accuracy and
power efficiency.

8. Robotics
Robots have been in existence for more than a century, but with the introduction of
AI, it has provided robots an ability to think on their own. It has started interacting,
performing tasks that would otherwise require human intervention. This has been
made possible through AI technologies, such as natural language processing, com-
puter vision, advanced motion control, etc.
80 Sourabh Sharma, Poonam Chaudhary

9. Navigations
Navigation was never as efficient as it is now in this modern world. Cab and logistics
companies are actively utilizing AI for lane detection, road and traffic analysis, and
optimum route selection.
Researchers are also working on developing AI models that provide optimum
routes with the most efficient time, keeping in mind the safety indicators.

4.3 Machine learning

ML is a subset of AI, where an algorithm learns from preexisting/historical data to
create a program that predicts the output of new/ unseen data [9]. The majority of ML
algorithms may be categorized into three: supervised learning, unsupervised learn-
ing, and reinforcement learning. Figure 4.2 represents the algorithms and techniques
for machine learning.

Machine
learning

Supervised unsupervised
learning learning

Classification Regression Clustering

Figure 4.2: Algorithms and techniques for machine learning.

4.3.1 Supervised machine learning

In supervised learning (SL), a variety of techniques are employed, including neural

networks, support vector machines, decision trees, and Bayesian networks [2–4]. A
key reason to consider the SL a precision model is that the model predicts the value of
one dataset based on the values of the other datasets. Based on pattern recognition,
predictive models prove to be phenomenal in dealing with uncertainty, which is a big
problem for off-road traffic. Algorithms are generally applied to produce models
using input data and known results of that data, i.e., the outputs, and to train that
model to make practical predictions regarding the response to new data [1]. An SL
model is shown in Figure 4.3 in its exact working state.
 Chapter 4 Machine learning and deep learning 81

Training data
(contains both input & output
& pattern in them)

Algorithm

New Input Data Classifier Prediction (Final output)

Figure 4.3: Model of supervised learning.

4.3.2 Using unsupervised machine learning

The most essential quality of Unsupervised Learning [6] is that it will master and learn
linked to the data on its own, i.e., learning from the environment. Three main tasks in-
volve using unsupervised learning models: clustering, association, and reduction of di-
mensionality. Clustering, which is predominantly employed in the electrical signal’s
framework, and density-based k-means algorithms are among the most commonly
used. Clustering, which employs a metric or attribute to show similarity and fabricates
dimensionality reduction to observe and evaluate high-dimensional data, is one of the
most widely used approaches. This can also be referred to as feature selection [16].

4.3.3 Reinforcement learning

Through frequent contact with a dynamic environment that either rewards or penalizes
the agent’s action, Reinforcement Learning (RL) [11] is an ML paradigm, where an agent
learns the best course of action for a particular task. However, the recent success of DL
algorithms has rekindled interest in the topic of RL among scientists, who are now ef-
fectively using this to solve extremely difficult problems that were previously thought
to be intractable. The rise of AI, which may soon surpass human intelligence, has gar-
nered attention globally as a result of events like artificial agents like AlphaGo defeating
world champion Lee Sedol or IBM Watson winning the game of Jeopardy.
Building such intelligent systems that can learn from their experiences over time
requires the use of the RL paradigm. Robotics, healthcare [17, 18], recommender systems,
data centers, smart grids, financial markets, and transportation are all areas where rein-
forcement algorithms are being used more and more, these days.
82 Sourabh Sharma, Poonam Chaudhary

4.3.4 Deep learning

Neural Network is a subset of ML that aims to mimic the behavior that a human brain
shows during learning. It uses a network of connected nodes called Neurons that per-
form some computations on received information and pass on their outputs to other
neurons as inputs. Artificial Neural Network is the oldest type of Neural Network,
where one or more fully connected hidden layers are present between the input and
the output [8]. These layers are responsible for learning from input data, using back
propagation techniques.

4.4 Humans and AI

AI models have changed the very dynamics of the virtual world– on the one hand it
has proved really handy and it has acted as a catalyst in the field of industry automa-
tion; on the other hand, it has sent turbulence across the sectors that it might reduce
employment opportunities replacing humans. Furthermore, AI models are not fully
perfect; there are always chances of errors from highly trained models, because ma-
chines lack the emotional component and ethical wisdom that humans possess [9].
To clear these misconceptions, we need to understand that AI models are just an
automation asset to utilize human resources more efficiently, so that we can use them
on more sophisticated tasks rather than underutilizing them on repetitive tasks. This
would enable human society to solve more complex problems that our society and
world are coping with and which require more human attention and wisdom.
Secondly, we cannot allow AI models to work independently, as they lack emo-
tional and ethical wisdom that only humans possess. Thus, we need smart systems
that have human inputs at certain points; and specially, there should be specific
checkpoints where functioning of models should be cross-checked by a team of hu-
mans to ensure sustainability of AI models to ensure they do not cause any harm.
For instance, a team of sustainers ensures the ethical dilemma of the data that is
being fed to the machine and another team of sustainers anticipates safety and harm
during the functioning. We will read more about this in ethics and AI. Concluding the
discussion, AI is not just a standalone making decisions themed from some fiction
movie, instead, it is a collaborative concept where humans and machines assist each
other to save resources and time and solve complex social, health, security, and eco-
nomic problems more efficiently.
 Chapter 4 Machine learning and deep learning 83

4.5 Ethics and AI

AI has led a revolution in the world of computing. It has enabled computer tasks that
could have been only imagined or written in comics [10]. The emergence of AI tech-
nology has resulted in saving a lot of time and computation power. Combining all
these factors along with higher different evaluation metrics achieved by the SOTA
(state-of-the-art) models have made the computer systems smarter and reduced inter-
vention of humans in those tasks.
These achievements are really astonishing, but an ethical code of conduct is the
most essential pillar of any technology. At this very point, ethics of AI comes into the
picture. Over a period of time, multiple stakeholders from various backgrounds have
raised legal and social issues pertaining to AI that might result in fatal undesirable
outcomes. AI models, like other statistics models, are prone to biases – they might be
discriminating in nature, they can be manipulative to change our behavior, and
might compromise our security.
AI ethics can be described as systemizing, defending, and recommending con-
cepts of right and wrong conduct to foster responsible use of AI.
AI ethics can be classified into three subfields:
– Meta Ethics: It is related to the study of meaning of ethical concepts, existence of
ontology, and epistemology.
– Normative Ethics: This field deals with the use of conceptual knowledge in prac-
tice for determining a moral philosophy.
– Applied Ethics: This field concerns the action of an ideal person acting in situations
of ethical dilemma.

These three subfields empower an AI practitioner with the wisdom to design better
and more efficient models, which are in compliance with the ethical and moral values
of society.

4.6 Conclusion
ML and DL algorithms have evolved over the years from conceptual algorithms en-
closed on pieces of paper to real-world applications helping the world achieve re-
markable results. With the introduction of quantum computing hardware, which
relies on laws of physics from quantum theory, models trained on these hardware are
supposed to outperform the models trained on existing computing resources. How-
ever, along with the growth of technology, wise and judicious use for the betterment
of society is an equally important facet in the world of emerging technologies.
84 Sourabh Sharma, Poonam Chaudhary

References
[1] L. Fanti, D. Guarascio, M. Moggi, From Heron of Alexandria to Amazon’s Alexa: A stylized history of AI
and its impact on business models, organization and work. J. Ind. Bus. Econ., 49(3), 2022, 409–440.
[2] L. Diop, S. Samadianfard, Annual rainfall forecasting using hybrid artificial intelligence model:
Integration of multilayer perceptron with whale optimization algorithm, Water Resour. Manag.
https://doi.org/10.1007/s11269-019-02473-8,2020.
[3] H. Wang, H. Zhang, K. Wang, C. Zhang, C. Yin, X. Kang, Off-road path planning based on improved
Ant Colony Algorithm, Wireless Pers. Commun., 2018, doi: 10.1007/s11277-017-5229-5.
[4] W. Aidan, A brief introduction to supervised learning, Towards Data Sci. 2019.
[5] P. Kaur, S. Harnal, R. Tiwari, S. Upadhyay, S. Bhatia, A. Mashat, A. M. Alabdali, Recognition of leaf disease
using hybrid convolutional neural network by applying feature reduction, Sensors. 22(2), 2022, 575.
[6] J. E. Van Engelen, H. H. Hoos, A survey on semi-supervised learning, Mach. Learn. 109(2), 2019, 373–440.
[7] M. Abouyoussef, S. Bhatia, P. Chaudhary, S. Sharma, M. Ismail, Blockchain-enabled online
diagnostic platform of suspected patients of COVID-19 like pandemics, IEEE Internet Things Mag. 4
(4), 2021, 94–99.
[8] P. Chaudhary, V. Kalra, S. Sharma, “A Hybrid Machine Learning Approach for Customer
Segmentation Uszing RFM Analysis,” International Conference on Artificial Intelligence and
Sustainable Engineering, 2022, pp. 87–100. Springer, Singapore.
[9] N. Yildirim, A. Kass, T. Tung, C. Upton, D. Costello, R. Giusti, S. Lacin, S. Lovic, J. M. O’Neill,
R. O. R. Meehan, E. Ó Loideáin, “How Experienced Designers of Enterprise Applications Engage AI as
a Design Material,” CHI Conference on Human Factors in Computing Systems, 2022, April, pp. 1–13.
[10] L. Zhu, X. Xu, Q. Lu, G. Governatori, J. Whittle, AI and Ethics – Operationalizing Responsible AI, in:
Humanity Driven AI, Springer, Cham, 2022, pp. 15–33.
[11] S. Safdar, M. Rizwan, T. R. Gadekallu, A. R. Javed, M. K. I. Rahmani, K. Jawad, S. Bhatia, Bio-imaging-
based machine learning algorithm for breast cancer detection, Diagnostics. 12(5), 2022, 1134.
[12] V. Van Roy, D. Vertesy, G. Damioli, AI and robotics innovation. Handbook of labor, human resources
and population economics, 2020, pp. 1–35.
[13] I. H. Sarker, Ai-based modeling: Techniques, applications and research issues towards automation,
intelligent and smart systems, SN Comput. Sci. 3(2), 2022, 1–20.
[14] S. Bhatia., A. K. Dubey, R. Chhikara, P. Chaudhary, A. Kumar, Intelligent Healthcare, Springer
International Publishing, New York, US, 2021.
[15] G. Alam, I. Ihsanullah, M. Naushad, M. Sillanpää, Applications of artificial intelligence in water
treatment for optimization and automation of adsorption processes: Recent advances and
prospects, Chem. Eng. J. 427, 2022, 130011.
[16] S. R. Swarna, A. Kumar, P. Dixit, T. V. M. Sairam, “Parkinson’s Disease Prediction using Adaptive
Quantum Computing,” 2021 Third International Conference on Intelligent Communication
Technologies and Virtual Mobile Networks (ICICV), 2021, pp. 1396–1401, doi: 10.1109/
ICICV50876.2021.9388628.
[17] G. Sasubilli, A. Kumar, “Machine Learning and Big Data Implementation on Health Care data,” 4th
International Conference on Intelligent Computing and Control Systems (ICICCS), 2020,
pp. 859–864, doi: 10.1109/ICICCS48265.2020.9120906.
[18] S. M. Sasubilli, A. Kumar, V. Dutt, “Improving Health Care by Help of Internet of Things and Bigdata
Analytics and Cloud Computing,” 2020 International Conference on Advances in Computing and
Communication Engineering (ICACCE), 2020, pp. 1–4, doi: 10.1109/ICACCE49060.2020.9155042.