Part B : Subject Specific Skills

Unit 1 - REVISITING AI PROJECT CYCLE & ETHICAL FRAMEWORK

Part – 1 – AI Project Cycle

AI Project Cycle
The AI project cycle is a structured, step-by-step framework used to develop and deploy artificial
intelligence (AI) solutions to solve specific problems. It is also known as the AI development lifecycle. The AI
Project Cycle mainly has 6 stages:

1. Problem Scoping - Problem scoping is the first stage of the AI Project Cycle. Problem Scoping refers to
understanding a problem, finding out various factors which affect the problem, define the goal or aim of
the project.
Trying to see or define what is to be done to solve a problem is called problem scoping and once
it is defined, then it is called the problem scope. A problem scope is mutual understanding of all
stakeholders about what is to be done to solve that problem.

2. Data acquisition: Data acquisition is the second stage of the AI Project Cycle. This stage focuses on
Collecting and compiling the relevant data to train the AI system. Data is the fuel for AI Algorithms.
When you want AI systems to make predictions, you need to use data to train them. Data collected
should be from reliable and authentic sources.
Some sources of data include Surveys, Web Scraping, Sensors, Cameras, Observations, and
Application Programming Interfaces (APIs)
Since the data you collect would be in large quantities, you can try to give it a visual image of
different types of representations like graphs, databases, flow charts, maps, etc. This makes it easier for
you to interpret the patterns which your acquired data follows.

3. Data exploration: Data exploration is the third stage of the AI Project Cycle. This stage involves
analyzing the data collected to interpret patterns, trends and relationships. Arranging the information in
a proper format or layout such as graphs, databases, flowcharts, and maps to easily identify patterns and
gain a clearer understanding.

4. Modelling: Modelling is the fourth stage of the AI Project Cycle. After exploring the patterns, you can
decide upon the type of model you would build to achieve the goal. AI Modelling involves developing
an algorithm that can draw conclusions or make predictions based on available data.
 The chosen model should be capable of learning from the data and making accurate predictions.

5. Evaluation: Evaluation is the fifth stage of the AI Project Cycle. This stage includes comparing the
model’s result with expected outcomes to access accuracy, reliability and overall performance. If the
model’s performance does not meet expectations, adjustments, retraining or fine tuning may be required
to enhance its effectiveness.
Common evaluation metrics include accuracy, precision, recall and F1 score depending on
the task.

6. Deployment : Deployment is the last stage of the AI Project Cycle. It involves integrating the trained
and evaluated AI model into a real-world application, system or product, allowing it to start making
predictions, decisions, or automating tasks based on incoming data.
This model is made accessible to users or other systems to achieve the project’s objectives,
ensuring it delivers maximum value and impact to users and stakeholders.

Subjective Type Questions

A. Very Short answer type questions.
1. What do you mean by the term project?
Any planned activity to be finished over a period of time to achieve set goals is called project.

2. List the stages of an AI Project Cycle.

Problem scoping, Data acquisition, Data exploration, Modelling, Evaluation.

3. What do you mean by stakeholder?

People affected by the solution of the problem directly or indirectly are called stakeholders.

4. What is a problem statement?

Describing the exact problem in concise and complete form is called problem statement.

5. What is a goal?
The achievable solution to the problem as per the plan is called a goal.

B. Short Answer Type Questions

1. What is Project Cycle ?

Project Cycle is a step by step process to solve the problems using proven scientific
methods and drawing the inference about it.
Creating a birthday card

1. Checking the factors like budget,etc Which will help us decide the next steps and understanding
the Project.
2. Acquiring data from different sources like online, with friends etc for Designs and ideas.
3. Making a list of the gathered data.
4. Creating or Modelling a card on the basis of the data collected.
5. Showing it to Parents or cousins to Let them check it or evaluate it.
 2. Briefly explain the first W of 4W framework in problem scoping.
The major scoping activities are described as part of 4 important parameters defined by 4W
framework. 4Ws stand for: Who, What, Where and Why

Who : Primarily, we should be familiar with the people who are facing the problem and the people who
will be affected by the solution of the problem directly or indirectly. All such people are called
stakeholders. The stakeholders include people in various capacities such as operators of the system for
which problem needs to be solved, users of the system, investors, managers and owner of the system etc.

3. How can we identify a problem?

The scope of the problem cannot be defined until the problem is understood completely and
correctly. Getting familiar with all aspects of the problem is the prerequisite to defining problem scope.
A problem which is identified and understood could be described in writing. This is called problem
statement. The problem statement is short and includes the problem description.

4. List 2 examples of specific goals.

• Develop the handwriting analyser within 60 days to analyse 500 samples in 1 minute.
• Create the burglar alarm system in 30 days that raises alarm of 120 decibel hearable within 500
meters radius.

5. What do you mean by the context of the problem?

The boundary of the problem area is defined by the context. What is the problem about? Are
there any proposed solutions? Who are the stakeholders? Who will work upon the problem? Who is
going to benefit from the solution?

6. What is fourth W for in 4W framework of problem scoping?

Why is the 4th W of 4W framework? This question answers the rationale of the solution. It
describes the benefits to be drawn from the implemented solution to the problem at hand. It also helps
you describe how valuable will be the solution to the stakeholders. The answer to this question must
inform the stakeholders how the situation will be improved after implementing the solution.

C. Long answer type questions.

1. Briefly describe the 4W framework in problem scoping.
The major scoping activities are part of 4 important parameters defined by 4W framework -
Who, What, Where and Why.

Who : Primarily, we should be familiar with the people who are facing the problem and the people who
will be affected by the solution of the problem directly or indirectly. All such people are called
stakeholders. The stakeholders include people in various capacities such as operators of the system for
which problem needs to be solved, users of the system, investors, managers and owner of the system etc.

What : The scope of the problem cannot be defined until the problem is understood completely and
correctly. It refers to what the problem is and what you know about the problem. What is the nature of the
problem? Can it be explained simply? How do you know it’s a problem? What is the evidence to support
that the problem exists? What solutions are possible in this situation? etc. At this stage, you need to
determine the exact nature of the problem. Once the problem is understood, it becomes easier to set the
goals to solve the problem.
 Where : This question defines the context or situation or location of the problem. It shows the exact
situation, area or boundary where the problem is occurring. It clearly shows you when and where exactly
the real problem arises and helps you pinpoint the affected area of the system.

Why : This question answers the rationale of the solution. It refers to the reason we need to solve the
problem, the benefits which the stakeholders would get from the solution and how would it benefit them
as well as the society, what are the benefits to the stakeholders after solving the problem.

2. How are goals and problem statement related to each other. Explain with an example.

Problem statement describes the stakeholders, problem and the proposed solution to the problem.
The goals are translated into the proposed solution to the problem in the problem statement. Goals cover
the entire purpose of the problem-scoping exercise. Without clear goals, problem statement cannot be
completed. For example, to develop a system that predicts the behaviour of the customers whether they
shall buy the new product or not, the problem statement looks like this:

3. Apply AI project Cycle to come up with an innovative solution of managing library books in your
school-the whole process of search, issuing and returns management.

Using latest advances in artificial intelligence (AI) in a library to meet the needs of both the users and the
library staff, AI can be useful in many ways such as:

(a) To facilitate the process of reserving a book and elastic searching for users or books using a QR code.

(b) Using machine learning (ML) technologies, all data can be analyzed, and librarians can observe trends in
readers’ choices and create an automated system of recommendations and acquire new books which may be
of interest to readers or give feedback to the publishing houses.

(c) Each book in the library can have embedded Radio Frequency Identification (RFID) tags to quickly scan
the book using a wireless handheld RFID scanner.

(d) Smart shelves containing several RFID antennas can help in maintaining issued or returned books or
even missing books.
 Part B : Subject Specific Skills
Unit 1 - REVISITING AI PROJECT CYCLE & ETHICAL FRAMEWORK
Part – 2 - Domains of AI

Domains of AI
In every technology involving computers, the underlying concept is input-processing
output. In AI, this can be rewritten as data-analysis-prediction. There are three domains of AI :
a. Statistical Data (Data Science)
b. Computer Vision
c. Natural Language Processing

1. List main capabilities of AI.

Main capabilities of AI are
i. Analysis of simple and complex immense data,
ii. Smart search
iii. Natural language processing.

2. What is Statistical Data/Data science/Big data? Give an example of it.

Data is the key of the three domains of AI. Data sciences is a domain of AI related to data systems and
processes, in which the system collects numerous data, maintains data sets and derives meaning/sense
out of them. The information extracted through data science can be used to make a decision about it.
OR
Data science is the field of study that combines domain expertise, programming skills, and knowledge of
mathematics and statistics to extract meaningful insights from data.
OR
Data Sciences, it is a concept to unify statistics, data analysis, machine learning and their related
methods in order to understand and analyses actual phenomena with data.

Some application of Statistical Data are - Price Comparison Websites, Website Recommendations,
Fraud and Risk detection, Internet search, Personalized healthcare recommendations, Optimizing Traffic
routes in real-time, image tagging.

Some possible variables in the above examples are :

o How many customers checked the product once, twice and thrice?
o How many bought it after one visit, two or more visits?
o How many customers recommended the product to others?
o How many new customers returned the product?
o What are the product reviews and feedbacks?
o What are the new features upgrades in the product?
o What is the comparative price of the product over the years?

Where do we collect data from?

Data can be collected from various sources like –
• Surveys
• Sensors
• Observations
• Web scrapping (Internet)
• Interviews
 • Documents and records.
• Oral histories

3. What is Computer Vision? Give an example of it.

Computer Vision, abbreviated as CV, is a domain of AI that depicts the capability of a machine
to get and analyze visual information and afterwards predict some decisions about it. The entire process
involves image acquiring, screening, analyzing, identifying and extracting information.
OR
Computer vision is an interdisciplinary field that deals with how computers can be made to gain
high-level understanding from digital images or videos.
OR
The Computer Vision domain of Artificial Intelligence enables machines to see through images or
visual data, process and analyze them on the basis of algorithms and methods in order to analyze actual
phenomena with images.

For Example: - Agricultural Monitoring, Surveillance Systems, Self-Driving cars / Automatic

Cars, Face Lock in Smartphones

4. What is Natural Language Processing? Give an example of it.

Natural Language Processing, or NLP, is the sub-field of AI that helps computers to understand
and process human languages like English or Hindi to analyze and derive it meaning. Natural language
refers to language that is spoken and written by people, and natural language processing (NLP) attempts
to extract information from the spoken and written word using algorithms.
In NLP, we teach machines how to understand and communicate in human language. Natural
language refers to speech analysis in both audible speeches, as well as text of a language. NLP systems
capture meaning from an input of words (sentences, paragraphs, pages, etc.)

Different algorithms of NLP can be used in assessing subjective type exam. For example, tokenizing
words and sentences, Parts of Speech tagging, Lemmatizing words and Word netting to evaluate the
subjective answer. Besides, NLP algorithm also provides the semantic meaning in context.

For Example:
a. Email filters or Spam Filters
b. Sentiment Analysis
c. Machine Translation - Google translate is one such application of NLP.
d. Text Summarisation
e. Interactive Voice Response Systems (IVRS) use NLP.
f. Virtual Assitants / Smart assistants: Google Assistant, Apple’s Siri and Amazon’s Alexa use
NLP techniques to interact intelligently.

NLP works on two aspect – Syntax and Semantics of the language:

a. Syntax: Syntax refers to the grammer of the language and sentence formation. This is easier
to apply since a language has a structured set of rules.
b. Semantics: Semantics refers to the interpretations of the meaning carried by the words which
depend on the context. This is a big challenge since it involves figuring out the context, tone,
emotions and expressions carried by the voice.
 e.g. NLP can help in sentiment analysis that can determine the action of the employees while
clustering models help to extract patterns in the employees’ responses while calling or chatting
with customers at a call center.

5. What is NLP? How does it help in education and customer support field?
Natural Language Processing (NLP) is the technique used in AI field to understand and process
speech and voice. NLP is useful in education field to develop interactive courses and assessments,
voice-based research systems, voice command-based educational systems, multi-lingual courses and
smart teaching in classrooms etc. In customer support field NLP is used as smart voice response systems
and chatbots which understand what has been spoken. They answer structured questions and understand
the intent of the customer faster and accurate.

6. Briefly list the ways in which AI can help in E-Commerce industry?

E-Commerce industry has immense data for AI systems to process and perform useful forecasts and
predictions. They are used to recommend ad promote products through digital marketing and develop
better customer relations. AI can be used to create better shopping experience for customers and address
to their problems faster and accurately.

7. How do NLP, Data Vision and Machine Learning help in research and development field?
Research and development field is almost an endless ocean of data in any field. AI can help in
smart search based on text, voice and image, categorise and compile the search results and information,
summarise the information. Machine learning and deep learning systems can be used to create 3-D
designs.

8. What do you understand by Machine Learning?

Machine Learning is a subset of Artificial Intelligence which enables machines to improve at

tasks with experience (data). The intention of Machine Learning is to enable machines to learn by
themselves using the provided data and make accurate Predictions/ Decisions. OR Machine learning
focuses on the development of computer programs that can access data and use it to learn for
themselves. OR Machine learning is a data analytics technique that teaches computers to do what
comes naturally to humans and animals: learn from experience.

9. What do you understand by Deep Learning?

Deep Learning is the most advanced form of Artificial Intelligence. In Deep Learning, the
machine is trained with huge amounts of data which helps it in training itself around the data. Such
machines are intelligent enough to develop algorithms for themselves. OR Deep learning is an
artificial intelligence (AI) function that imitates the workings of the human brain in processing data
and creating patterns for use in decision making. OR Deep learning is a subset of machine learning
where artificial neural networks, algorithms inspired by the human brain, learn from large amounts
of data.

10. List four real life examples of enormous data generation.

Four real-life examples of enormous data generation are
a. Customer data on an E-commerce portal,
b. Financial transactions on a bank website,
c. Medication data of patients in a large hospital
d. Pollution data of a state.
11. Why is it necessary for an AI system to be fed with enormous amount of data to learn its task?
An AI algorithm needs to be trained with data to make it capable for analysing data in future
and identify trends and patterns then doing forecasts. Such data used to train an AI algorithm is
called testing data. Thus, performance of AI system is based primarily on the bulk of data. Higher
the amount of data, better would be the training of AI algorithm. This is the reason enormous
amount of rich data (with a variety of variables) is required to train AI algorithms.

12. How are three domains of AI related together?

Be it computer vision or NLP, data is the common domain that is required for learning by AI
algorithms. If that data is in the form of images and visuals then it is Computer Vision and if the data
is in the form of spoken language or audio then it relates to NLP. This way the 3 domains – data,
Computer Vision and Natural Language Processing are related together. For example, an AI
algorithm creating a summary of a news clip.

13. Briefly list any two examples of Computer Visions and NLP integration.
The 2 examples of Computer Vision are object identification in an image and face
recognition.
The 2 examples of NLP are voice based chatbots and question-answer based voice operated
assistant. Other examples are Alexa, Cortana and Siri.
 Part B : Subject Specific Skills
Unit 1: AI Project Cycle and Ethical Framework
Part 3: Ethical Framework

Learning Outcomes
a. Understand the Ethical Issues related to AI.

Watch the video : AI for Good https://www.youtube.com/watch?v=vgUWKXVvO9Q

Note – Write first 10 questions in your copy.

1. What do you mean by ethics? Explain briefly why ethics is important with AI development?

Ans : The principles and morals that govern our behaviour are called ethics. They serve as a guideline for us to
distinguish between right and wrong. Artificial intelligence involves faster processing and a remarkable impact (both
positive and negative) on industries, society and environment. It has the ability to go out of control if not monitored
properly through stringent policies and great care. Unethical use of AI may lead to great many threats to human
rights, equality issues (biases), life threats, social conflict and negative effect in the application area instead of
providing intended benefit. This is the reason addressing ethical issues and devising their preventive plans are
important.

2. List any 5 ethical characteristics of good AI?

Ans : Five ethical characteristics of good AI are enriching human life, working towards equity, easing conflicts and
promoting harmony, addressing un-addressable problems and addressing environmental abuse.

3. Briefly explain transparency and bias issue related to AI?

Ans : Think of a self-learning and evolving AI system. There should be enough and appropriate guidelines and
process to determine the ethics of action taken by such an AI system. The accuracy and suitability of data used to
train it needs stringent checks. Ethical guidelines need to be in place to ensure that AI enabled system is taking
decision and serving in a transparent manner avoiding any sort of bias on any basis such as gender, race, ethnicity,
region, financial and social status etc. For such biases, proper accountability guidelines need to be devised. This all
will ensure an ethical AI-based service system.

4. List any 5 major ethical challenges related to AI?

Ans : Four major ethical challenges related to AI are accountability, biases, privacy and human values.

5. What do you understand by the term Frameworks?

Ans: Frameworks are a set of steps that help us in solving problems. It provides a step-by-step guide for solving
problems in an organized manner. Moreover, frameworks offer a structured approach to problem-solving, ensuring
that all relevant factors and considerations are taken into account. Additionally, they serve as a common language
for communication and collaboration, facilitating the sharing of best practices and promoting consistency in
problem- solving methodologies.
6. What is Ethical Frameworks?

Ans: Ethical frameworks are frameworks which help us ensure that the choices we make do not cause unintended
harm. Furthermore, ethical frameworks provide a systematic approach to navigating complex moral dilemmas by
considering various ethical principles and perspectives. By utilizing ethical frameworks, individuals and organizations
can make wellinformed decisions that align with their values and promote positive outcomes for all stakeholders
involved.

7. What are the types of Ethical Frameworks?

Ans: The various types of ethical frameworks are classified as follows :

Ethical frameworks for AI can be categorized into two main types: sector-based and value-based frameworks.

1. Sector-based Frameworks: These are frameworks tailored to specific sectors or industries. In the context of AI,
one common sector-based framework is Bioethics, which focuses on ethical considerations in healthcare. It
addresses issues such as patient privacy, data security, and the ethical use of AI in medical decision-making. Sector-
based ethical frameworks may also apply to domains such as finance, education, transportation, agriculture,
governance, and law enforcement.

2. Value-based Frameworks: Value-based frameworks focus on fundamental ethical principles and values guiding
decisionmaking. It reflects the different moral philosophies that inform ethical reasoning. Value-based frameworks
are concerned with assessing the moral worth of actions and guiding ethical behaviour. They can be further
classified into three categories:

i. Rights-based: Prioritizes the protection of human rights and dignity, valuing human life over other considerations.
It emphasizes the importance of respecting individual autonomy, dignity, and freedoms. In the context of AI, this
could involve ensuring that AI systems do not violate human rights or discriminate against certain groups.

ii. Utility-based: Evaluates actions based on the principle of maximizing utility or overall good, aiming to achieve
outcomes that offer the greatest benefit and minimize harm. It seeks to maximize overall utility or benefit for the
greatest number of people. In AI, this might involve weighing the potential benefits of AI applications against the
risks they pose to society, such as job displacement or privacy concerns.

iii. Virtue-based: This framework focuses on the character and intentions of the individuals involved in decision-
making. It asks whether the actions of individuals or organizations align with virtuous principles such as honesty,
compassion, and integrity. In the context of AI, virtue ethics could involve considering whether developers, users,
and regulators uphold ethical values throughout the AI lifecycle.
8. What is Bioethics?

Ans: Bioethics is an ethical framework used in healthcare and life sciences. It deals with ethical issues related to
health, medicine, and biological sciences, ensuring that AI applications in healthcare adhere to ethical standards and
considerations.

Principles of bioethics: • Respect for Autonomy. • Do not harm. • Ensure maximum benefit for all. • Give
justice.

• “Non-maleficence” refers to the ethical principle of avoiding causing harm or negative consequences. It
emphasizes the obligation to minimize harm as much as possible and prioritize actions that prevent harm to
individuals, communities, or the environment.
• “Maleficence” refers to the concept of intentionally causing harm or wrongdoing.
• “Beneficence" refers to the ethical principle of promoting and maximizing the well-being and welfare of
individuals and society. It emphasizes taking actions that produce positive outcomes and contribute to the
overall good, ensuring that the greatest benefit is achieved for all stakeholders involved.

9. List the common types of bias that can occur in AI systems. How can they impact decision-making?
Ans. Common types of bias in AI systems include data bias, sampling bias, historical bias, and gender bias. These
biases can lead to unfair decisions, as algorithms may be trained on flawed or incomplete data, causing
inaccuracies in results. For example, a hiring algorithm biased against certain races can lead to unfair candidate
selection.

10. What is the role of AI Ethicists? Why are they important in the AI industry?
Ans. AI Ethicists study the ethical impact of AI and ensure that AI is used in a responsible and beneficial manner.
They are important to prevent biased algorithms and ensure fairness and transparency in AI systems.

11. State the main ethical concerns associated with AI with reference to job displacement, wealth inequality, AI
access, and AI bias.
Ans. The main ethical concerns include job displacement, wealth inequality, AI access, and AI bias. Job
displacement arises as automation may replace human jobs. Wealth inequality may result from AI-driven
companies profiting at the expense of employees. Not everyone has equal access to the benefits of AI, which
results in inequality. AI bias occurs when algorithms produce biased results due to biased training data, posing
risks to decision-making processes.

12. Explain any one example of AI bias.

 Ans. Gender bias is an example of AI bias. This bias can occur due to various reasons, including biased training
data, flawed algorithms, or the influence of societal biases present in the data used to train the AI system. For
example, a company once discovered that the algorithm they used for hiring employees was biased against
women. The algorithm was based on the number of resumes submitted over the past ten years, and since most
of the applicants were men, it favoured men over women.

13. In a brief paragraph, explain how a policy framework for good AI will help in establishing ethical AI.

Ans : A policy framework can help in establishing ethical AI by revising existing human rights policies in the light of
AI, considerations of impact of AI on economy and gross domestic product, people’s right to their data and privacy,
testing and benchmarking systems for measuring efficacy and effectiveness of AI systems, laws to address AI system
malfunction consequences and issues, devising accountability for functioning of AI systems, ensuring screening
policies to ensure authorised, qualified and expert individual involve in developing and handling AI ecosystems.

14. If a machine based on AI makes a decision with unintended consequences, who is responsible?

Ans : Apart from algorithms and data, researchers and engineers developing these systems are also responsible for
AI bias. Bias can creep in AI systems from Developers and Data.

15. An online search shows mostly male candidates when searched for ‘doctors’ and female candidates when
searched for ‘nurses’. Should the developer be held accountable for this bias?

Ans : If AI develops a certain bias against a gender, then the accountability will lie on how it was taught and trained
by the developer. The developer should check the facts from the government census before writing an algorithm to
create a search engine that produces data on potential discrimination.

16. A company decides to select candidates using an AI system for the role of ‘Head Chef’. What all biases are
likely to be present in this case?

Ans : It will exhibit bias against employment discrimination, gender discrimination and prejudice. In this case, the
data used to train the system reflects existing prejudices, stereotypes and other faulty societal assumptions. Using
data about ‘Head’ that includes only males would thereby preserve a real-world gender stereotype in the computer
system.

17. In a company, visitors enter their personal details in visitor’s book. However, the company shares this data
with some marketing companies. Do you think it is ethical? Why/Why not?

Ans : It is unethical when the company does not publicly state that the information would be used for commercial
purposes and does not take consent of the people. Customers should be informed specifically what their
information will be used for and how and by whom it will be used. It can be ethical if a company states in their
privacy policies that they collect your data and share it with third parties for improvement purposes.

18. A self-driving car is programmed to protect innocents from criminals. In case of accidents, it hits a murderer
on the road rather than a puppy. Should the developer be held accountable for this decision.

Ans : The developer is not seemingly accountable as the self-driving car is programmed to protect the innocent. But
it does not mean the criminal should not be punished. In case of accident, the car might be programmed to give first
priority to save the driver and if the probability of saving a driver increases, the car hits the human obstacles
(murderer) and not the puppy.
19. Some experts say robots for military combats will be more efficients. However, others challenge it on ethical
grounds. What are the ethical challenges to it.

Ans : Robots are highly useful in war for searching, tracking, safely diffusing bombs, and rescuing even in nuclear,
biological, radiological and chemical environments. They can be automated or operated remotely by soldiers. But
the use of automated robots raises ethical and risk concerns.

a. The first ethical concern involves the use of any technology-led military combat during war that
would violate existing international law covenants.
b. Secondly, currently robots are not as capable of distinguishing between civilians and enemy combatants. As
such, these robots have the potential to target and kill innocent civilians.

20. Human values are simple at its core. However, factors like culture, law, religion etc, make them complex. This
complexity is at the core of ethics of AI. Discuss with examples.

Ans : Human values are based on our experiences and the cultures we have lived in and achieving ethical,
trustworthy and profitable AI requires that deliberations be grounded in a scientific understanding of the relative
strengths and weaknesses of both machine intelligence and human cognition. The principle of AI ethics holds that AI
should be designed to help promote the well-being of people and the planet.

For example,

• AI created to help people overcome visual or hearing impairment,

• Chatbots that deliver Cognitive Behavioural Therapy (CBT) to help people in conditions such as low-level
depression and anxiety.
• Social robots that incorporate growth mindset to help children stay focused in learning environments.
Justice is another core ethical principle which is due to algorithm fairness that provides equality, diversity,
unbiasedness and treats people equally.

21. Do you consider it ethical to incorporate the use of AI for various jobs, given that it takes away human jobs
and magnifies human bias?

Ans : Automation is a tool and use of automation software is ethical. As automation becomes more proficient,
people worry more about jobs being taken over. An automated machine is there to help people rather than to
replace them. Automation can improve the working life and further an employee’s career by training them. If it is
treated as an assistant and not as a replacement, then no ethical problems will arise.

22. How do you think income would be shared if AI is used in place of human workforce?

Ans : Like all machines, AI is also owned by an individual or an entity. On its own, it doesn’t have any income and IP
is owned by someone. So, if the AI machine generates income, then that goes to the owner who can use it for the
development/ maintenance of the machines or the electricity it will consume.

23. AI will probably bring with it many health benefits and learning opportunities. How will these benefits be
made accessible and available to all the people in the society?

Ans : Learning opportunities: Various learning applications, tutoring programs and skill-based curriculum are being
developed and will become easily accessible to students using smartphones. These apps not only empower students
but also help teachers in upgrading themselves with current trends. Besides, such systems could be a boon for the
rural population.
Health-care benefits: Wearable devices are available for checking blood pressure, respiratory rate, oxygen levels,
pulse, blood pressure and body temperature

24. AI is a powerful tool in various fields, however, depending on how it is used, it can either be blessing or a
curse. Discuss with an example?

AI already plays an essential role in our day-to-day lives—from Google Search and Alexa to Netflix and Uber. One
of the most famous and closest examples is the self-driving car. AI can, with its immense power and ability to
learn, put an end to global poverty, find a cure for cancer and ensure food security. Advantages of Artificial
Intelligence:
(a) AI-powered machines ensure a higher accuracy than humans.
(b) Using AI, doctors not only check health risks and sideeffects of various medicines but AI-driven
machines lend a hand during surgeries, which consequently lead to high success ratio.
(c) We can access desired information from the internet using AI-powered voice assistants like Siri,
Cortana, Alexa, etc.

Disadvantages of Artificial Intelligence:

(a) Implementation of AI is expensive due to its complex algorithms and maintenance.

(b) AI-based applications cannot make decisions on their own and act only in those environments for
which they are programmed.
(c) AI-based machines can replace humans as they can perform repetitive tasks and labour-intensive
tasks more efficiently and are, therefore, a threat to human employment. AI could be a blessing for
the human race if perceived properly and would be more effective in future.
25. How can learning opportunities for AI be extended to all?

Ans : There are numerous applications of AI for education that are accessible to all through smart devices and
computers. AI is also allowing educators to assist students who are struggling and provide human interaction and
hands-on experience to them. AI is already driving some changes in classroom such as Flipped classroom model that
allows students to watch pre-recorded video lectures on their own while class time is reserved for more active
learning like discussions, projects and workshops. AI can enable self-paced personalized learning environment for
every learner

26. How will human beings ensure that they stay ahead of AI?

Ans : AI is everywhere. It is making life easier and more efficient in its ability to automate processes and procedures.
But as AI becomes more visible in the workplace, there is growing concern that it may take over and replace
employees by taking up various roles. Humans are the ones who have invented AI, but with AI becoming generative,
humans may not be able to directly control the evolution of AI.

Few ways in which humans can stay ahead in the game are:

(a) Learn new skills.

(b) Be creative and learn AI.
(c) Prepare to defend yourself before developing dangerous AI technologies.
(d) Don’t give machines so much power that they overtake you
 Part B : Subject Specific Skills
Unit 2 - ADVANCED CONCEPTS OF MODELLING IN AI
Part-1

Learning Objectives:
1. To familiarize students with supervised, unsupervised and reinforcement learning based
approach.
2. To introduce students to the neural network

1. What is Machine Learning?

Machine Learning (ML) Machine Learning, or ML, enables machines to improve at tasks with
experience. The machine learns from its mistakes and takes them into consideration in the next execution.
It improvises itself using its own experiences.

Input (past or historical data) is given to the ML model and the model generates output by learning from
the input data.
Here is an example which shows labelled images (every image is tagged either as apple or strawberry) are
given as input to the ML model. ML model learns from the input data to classify between apples and
strawberries and predicts the correct output as shown.

Examples of Machine Learning (ML)

Object Classification - Identifies and labels objects present withinan image or data point. It determines
the category an object belongs to.
Anomaly Detection - Anomaly detection helps us find the unexpected things hiding in our data. For
example, tracking your heart rate, and finding a sudden spike could be an anomaly, flagging a potential
issue.

2. What is Deep Learning (DL)?

Deep Learning, or DL, enables software to train itself to perform tasks with vast amounts of data. In deep
learning, the machine is trained with huge amounts of data which helps it into training itself around the
data. Such machines are intelligent enough to develop algorithms for themselves. Deep Learning is the
most advanced form of Artificial Intelligence out of these three. Following is the block diagram of deep
learning:

Input is given to an ANN, and after processing, the output is generated by the DL block.

Here is an example which shows pixels of a bird image given asinput to the DL Model and the model is
able to analyze and correctly predict that it is a bird using a deep learning algorithm (ANN).

Examples of Deep Learning (DL)

Object Identification - Object classification in deep learning tackles
the task of identifying and labeling objects within an image. It
essentially uses powerful algorithms to figure out what's ina picture
and categorize those things.

Digit Recognition - Digit recognition in deep learning tackles the

challenge of training computers to identify handwritten digits (0-
9) within images.

3. What is Data?
• Data is information in any form
• For e.g. A table with information about fruits is data
• Each row will contain information about different fruits
• Each fruit is described by certain features

4. What do you mean by Features?

• Columns of the tables are called features
• In the fruit dataset example, features may be name, color,
size, etc.
• Some features are special, they are called labels

5. What are Labels?

Data Labeling is the process of attaching meaning to data

• It depends on the context of the problem we are trying to
solve.
• For e.g. if we are trying to predict what fruit it is based on
the color of the fruit, then color is the feature, and fruit name is the label.
• Data can be of two types – Labeled and Unlabeled Labeled Data
• Labeled Data - Data to which some tag/label is attached. For e.g. Name, type, number, etc.
• Unlabeled Data - The raw form of data. Data to which no tag is attached.
6. What do you mean by a training data set?
The training data set is a collection of examples given to the model to analyze and learn. Just like how
a teacher teaches a topic to the class through a lot of examples and illustrations. Similarly, a set of
labeled data is used to train the AI model.

7. What do you mean by a testing data set?

The testing data set is used to test the accuracy of the model. Just like how a teacher takes a class test
related to a topic to evaluate the understanding level of students. Test is performed without labeled
data and then verify results with labels.

8. What is Modelling?
AI Modelling refers to developing algorithms, also called models which can be trained to get
intelligent outputs. i.e writing codes to make a machine artificially intelligent.
• The graphical representation makes the data
understandable for humans as we can discover
trends and patterns out of it.
• But when it comes to machine accessing and
analyzing data, it needs the data in the most
basic form of numbers (0s and 1s) and when it
comes to discovering patterns and trends in
data, the machine goes for mathematically
representation of the same.
• The ability to mathematically describe the
relationship between parameters is the heart of every AI model.
• Thus, whenever we talk about developing AI models, it is the mathematical approach towards
analyzing data which we refer to.

9. What are the Types of AI Models?

AI Models are classified into two parts :

a. Rule-Based Approach
• Rule Based Approach refers to the AI modelling where the relationship or patterns in data are
defined by the developer.
• Rule Based AI produces pre-defined outcomes that are based on a set of certain rules coded by
developers. The machine follows the rules or instructions mentioned by the developer, and
performs its tasks accordingly. These rules are coded into the system in the form of if-then-
else statements.
Example - Rule-based Chatbots are commonly used on websites to answer frequently asked
questions (FAQs) or provide basic customer support.

The two major components of rule-based artificial intelligence models are “a set of rules” and “a
set of facts”.
A rule-based system is like a human being born with fixed knowledge. The knowledge of
that human being doesn’t change over time. This implies that, when this human being encounters
a problem for which no rules have been designed, then this human gets stuck and so won’t be able
to solve the problem.
a. Rule-based AI models are deterministic
b. Rule-based artificial intelligence developer models are not scalable.
c. Rule-based AI models can operate with simple basic information and data.
d. Rule-based artificial intelligence systems are immutable objects.
 1. When to utilize rule-based models
a. Not planning for machine learning
b. Danger of error
c. Speedy outputs
2. When not to utilize rule-based models
a. Chances of addition of contradicting rule while adding new rules. It’s tough to add rules to
an already large knowledge base without introducing contradicting rules.
b. Upgrade and upkeep of such systems are time consuming and expensive.
c. Not versatile to be used in multiple domains since all problems cannot be defined by a
structured set of rules.

b. Learning-based Approach
• In this approach, random data is fed to the machine and it is left on the machine to figure
out patterns and trends out of it.
• This approach is followed when the data is unlabeled and too random for a human to make
sense out of it.
• Thus, the machine looks at the data, tries to extract similar features out of it and clusters
same datasets together.
• In the end as output, the machine tells us about the trends which it observed in the training
data.

For Example - Prediciting customer behavior, monitoring financial transaction for frauds, medical
diagonistic, legal research and advise etc. are the application areas for learning based approaches.
Machine Learning and Deep Learning both are Learning based approaches.

For example, A learning-based spam email filter is a computer program that automatically
identifies whether an incoming email is spam or not. Instead of being explicitly programmed with
rules for identifying spam, the filter learns from examples of labeled emails during a training phase.

During training, the filter is provided with a large dataset of emails, each labeled as either spam or
legitimate (non-spam). The filter analyzes the content and characteristics of these emails, such as
words used, sender information, and presence of attachments. Using machine learning algorithms,
the filter learns to recognize patterns that distinguish spam from legitimate emails.

Once trained, the filter can classify new incoming emails as spam or not spam based on the
patternsit learned. It continuously adapts and improvesits accuracy over time asit encounters new
examples. This learning-based approach allows the filter to effectively identify and filter out spam
emails, helping users manage their email inbox more efficiently.

MACHINE LEARNING

Machine learning is a subset of artificial intelligence (AI) which focuses on the use of data
and algorithms to imitate the way that humans learn, gradually improving its accuracy.
Machine learning, deep learning, and neural networks are all sub-fields of artificial
intelligence. However, deep learning is actually a sub-field of machine learning, and neural
networks is a sub-field of deep learning.

• Machine learning systems are probabilistic.

• Machine learning models utilize statistical rules rather than a deterministic approach.
• Machine learning systems can be easily scaled.
 • Machine learning systems require more data as compared to rule-based models. Machine
learning systems require full demographic data details.
• Machine learning models are mutable objects that enable enterprises to transform the data
or value by utilizing mutable coding languages such as java.

When to utilize machine learning models

a. Pure coding processing
b. Pace of change
c. Simple guidelines don’t apply

10. What are the types of Machine Learning Based methods?

Machine learning classifiers fall into three primary categories:

1. Supervised Machine Learning

• In a supervised learning model, the
dataset which is fed to the machine
is labelled to train algorithms so that
to classify data or predict outcomes
accurately.

• In other words we can say that the

dataset is known to the person who
is training the machine only then
he/she is able to label the data. A label is some information which can be used as a tag for
data.

• Supervised learning helps organizations solve for a variety of real-world problems at scale,
such as classifying spam in a separate folder from your inbox. Students get grades
according to the marks they secure in examinations. These grades are labels which
categorize the students according to their marks.

• There are two types of supervised learning models :

o Classification
o Regression

• Some methods used in supervised learning include neural networks, naïve bayes, linear
regression, logistic regression, random forest,
support vector machine (SVM), multi-class
classification and Decision Trees.
2. Unsupervised Machine Learning :

• An Unsupervised learning model

works on unlabeled dataset. This
means that the data which is fed to
the machine is random and there is
a possibility that the person who is
training the model does not have
any information regarding it.
• The unsupervised learning models
uses machine learning algorithms to
analyze and cluster unlabeled datasets i.e used to identify relationships, patterns and trends
out of the data which is fed into it. These algorithms discover hidden patterns or data
groupings without the need for human intervention.
• Its ability to discover similarities and differences in information make it the ideal solution
for exploratory data analysis, cross-selling strategies, customer segmentation, image and
pattern recognition.
• Algorithms used in unsupervised learning include neural networks, k-means clustering,
probabilistic clustering methods, and more.

3. Reinforcement Learning :
AI MODELS

1. Decision Tree (Rule Based Approach)

• Decision tree is the most powerful and popular tool for classification and prediction.

• Decision tree looks like a classic binary tree that propagates by answering simple questions
based on the answer “Yes” and “No” and it further split the tree into subtrees. A Decision Tree
is the inverted figure of a real tree where roots are at the top and and leaves at the bottom.

• In a Decision tree, every component of a decision tree is called a node. There are three nodes,
which are the Root Node, Decision Node and Leaf Node.

• Decision Tree is a Supervised learning

technique that can be used for both classification
and Regression problems, but mostly it is
preferred for solving Classification problems. It
is a tree-structured classifier, where internal
nodes represent the features of a dataset,
branches represent the decision rules and each
leaf node represents the outcome.

• In order to build a tree, we use the CART algorithm, which stands for Classification and
Regression Tree algorithm.

Decision Tree Terminologies

1. Root Node : Root node is from where the decision tree starts. It represents the entire dataset,
which further gets divided into two or more homogeneous sets.
2. Decision node : There are always two or more decision nodes in a tree. Each node specifies a
new condition that emerges from the previous condition. Decision nodes are the outcome of
either a decision taken on the condition in positive (Yes) or negative (No).
3. Leaf Node: Leaf nodes are the final output node, and the tree cannot be segregated further
after getting a leaf node. Leaves specify the final decision taken after considering all the
conditions.
4. Splitting: Splitting is the process of dividing the decision node/root node into sub-nodes
according to the given conditions.
5. Branches : A node is connected with the next node either through a positive branch (Yes) or
negative branch (No). A tree formed by splitting the tree.
6. Pruning: Pruning is the process of removing the unwanted branches from the tree.
 7. Parent/Child node: The root node of the tree is called the parent node, and other nodes are
called the child nodes.

Note: A decision tree can contain categorical data (YES/NO) as well as numeric data.

Advantages of the Decision Tree

• It is simple to understand as it follows the same process which a human follow while making
any decision in real-life.
• It can be very useful for solving decision-related problems. The logic behind the decision tree
can be easily understood because it shows a tree-like structure.
• It helps to think about all the possible outcomes for a problem.
• There is less requirement of data cleaning compared to other algorithms.

Disadvantages of the Decision Tree

• The decision tree contains lots of layers, which makes it complex.
• It may have an overfitting issue, which can be resolved using the Random Forest algorithm.
• For more class labels, the computational complexity of the decision tree may increase.

Decision Tree Example :

The following is a dataset comprising of 4 parameters which lead to the prediction of whether
an Elephant would be spotted or not. The parameters which affect the prediction are :
Outlook, Temperature, Humidity and Wind. Draw a Decision Tree for this dataset.
2. Classification Model
• The Classification algorithm is a Supervised Learning technique that is used to identify the
category of new observations on the basis of training data. In
Classification, a program learns from the given dataset or
observations and then classifies new observation into a
number of classes or groups. Such as, Yes or No, 0 or 1,
Spam or Not Spam, cat or dog, etc.

• The output variable of Classification is a category, not a value,

such as "Green or Blue", "fruit or animal", etc.

• This model works where the data is classified according to

the labels. For example, in the grading system, students are classified on the basis of the
grades they obtain with respect to their marks in the examination. This model works on
discrete dataset which means the data need not be continuous.

• The algorithm which implements the classification on a dataset is known as a classifier. There
are two types of Classifications:
a. Binary Classifier: If the classification problem has only two possible outcomes, then
it is called as Binary Classifier.
Examples: YES or NO, MALE or FEMALE, SPAM or NOT SPAM, CAT or DOG,
etc.
b. Multi-class Classifier: If a classification problem has more than two outcomes, then it
is called as Multi-class Classifier.
Example: Classifications of types of crops, Classification of types of music.

• Types of Classification Model

a. Logistic Regression
b. Naïve Bayes
c. Stohastic Gradient Descent
d. Support Vector Machine

3. Regression Model
• This model works on continuous data. For example, if you wish
to predict your next salary, then you would put in the data of
your previous salary, any increments, etc., and would train the
model. Here, the data which has been fed to the machine is
continuous.
 • A regression problem is when the output variable is a real or continuous value, such as
“salary” or “weight”. It tries to fit data with the best hyper-plane which goes through the
points.

• Types of Regression Model :

a. Linear Regression
b. Stepwise Regression
c. Ridge Regression
d. Lasso Regression
e. Polynomial Regression

Note : The difference between the two tasks is the fact that the dependent attribute is
numerical for regression and categorical for classification.

4. Clustering Model
• This model refers to the unsupervised learning algorithm which
can cluster the unknown data according to the patterns or trends
identified out of it. The patterns observed might be the ones
which are known to the developer or it might even come up with
some unique patterns out of it.

• Clustering or cluster analysis is a machine learning technique,

which groups the unlabelled dataset. It can be defined as "A way
of grouping the data points into different clusters, consisting of
similar data points. The objects with the possible similarities remain in a group that has less
or no similarities with another group."

• It does it by finding some similar patterns in the unlabelled dataset such as shape, size, color,
behavior, etc., and divides them as per the presence and absence of those similar patterns.

• It is an unsupervised learning method, hence no supervision is provided to the algorithm, and

it deals with the unlabeled dataset.

• After applying this clustering technique, each cluster or group is provided with a cluster-ID.
ML system can use this id to simplify the processing of large and complex datasets.
The clustering technique is commonly used for statistical data analysis.

Note : Clustering is somewhere similar to the classification algorithm, but the difference is
the type of dataset that we are using. In classification, we work with the labeled data set,
whereas in clustering, we work with the unlabelled dataset.

• For Example: Let's understand the clustering technique with the real-world example of Mall:
When we visit any shopping mall, we can observe that the things with similar usage are
grouped together. Such as the t-shirts are grouped in one section, and trousers are at other
sections, similarly, at vegetable sections, apples, bananas, Mangoes, etc., are grouped in
separate sections, so that we can easily find out the things. The clustering technique also works
in the same way. Other examples of clustering are grouping documents according to the topic.
 • The clustering technique can be widely used in various tasks. Some most common uses of this
technique are:
o Market Segmentation
o Statistical data analysis
o Social network analysis
o Image segmentation
o Anomaly detection, etc.

• It is used by the Amazon in its

recommendation system to provide the
recommendations as per the past search of
products. Netflix also uses this technique to
recommend the movies and web-series to its
users as per the watch history.
The below diagram explains the working of the clustering algorithm. We can see the
different fruits are divided into several groups with similar properties.
Types of Clustering Methods
The clustering methods are broadly divided into Hard clustering (datapoint belongs to
only one group) and Soft Clustering (data points can belong to another group also). But there are
also other various approaches of Clustering exist. Below are the main clustering methods used in
Machine learning:
o Partitioning Clustering
o Density-Based Clustering
o Distribution Model-Based Clustering
o Hierarchical Clustering
o Fuzzy Clustering

Applications of Clustering
Below are some commonly known applications of clustering technique in Machine
Learning:
o In Identification of Cancer Cells: The clustering algorithms are widely used for the
identification of cancerous cells. It divides the cancerous and non-cancerous data sets into
different groups.
o In Search Engines: Search engines also work on the clustering technique. The search result
appears based on the closest object to the search query. It does it by grouping similar data
objects in one group that is far from the other dissimilar objects. The accurate result of a query
depends on the quality of the clustering algorithm used.
o Customer Segmentation: It is used in market research to segment the customers based on
their choice and preferences.
o In Biology: It is used in the biology stream to classify different species of plants and animals
using the image recognition technique.
o In Land Use: The clustering technique is used in identifying the area of similar lands use in
the GIS database. This can be very useful to find that for what purpose the particular land
should be used, that means for which purpose it is more suitable.

5. Dimensionality Reduction
• We humans are able to visualize upto 3-Dimensions only but according to a lot of theories and
algorithms, there are various entities which exist beyond 3-Dimensions.
• For example, in Natural language Processing, the words are considered to be N-Dimensional
entities. Which means that we cannot visualize them as they exist beyond our visualization
ability. Hence, to make sense out of it, we need to reduce their dimensions.
 Here, dimensionality reduction algorithm is used.
As we reduce the dimension of an entity, the
information which it contains starts getting distorted.
For example, if we have a ball in our hand, it is 3-
Dimensions right now. But if we click its picture, the
data transforms to 2-D as an image is a 2-
Dimensional entity.

Now, as soon as we reduce one dimension, at

least 50% of the information is lost as now we will
not know about the back of the ball. Whether the ball
was of same color at the back or not? Or was it just a hemisphere? If we reduce the dimensions
further, more and more information will get lost. Hence, to reduce the dimensions and still be
able to make sense out of the data, we use Dimensionality Reduction.