2 AI PROJECT CYCLE

SESSION 1: INTRODUCTION TO AI PROJECT CYCLE

Introduction
In your daily life, you must have heard the word ‘project’ innumerable mes. Do you know
what a project is? Let us discuss. A project can be defined as a set of inputs and outputs
required to achieve a predetermined or specific goal. It can also be defined as a series of tasks
that are needed to be performed in order to achieve a specific or predetermined goal within a
defined me limit. Some mes, there is a requirement to break the task of the whole project
into subtasks to achieve good produc vity in terms of me management, cost, etc. A project
can be simple as well as complex. Like the human life cycle, a project also follows a life cycle in
which various sequences of phases are available from beginning to end. The number and
sequence of the cycle depend upon the needs of an organiza on, nature of the project and its
area of applica on.
Even though every project has a definite start and end, the par cular objec ves, deliverables,
and ac vi es vary widely. The life cycle of a project gives us the ac on layout that has to be
performed in the project, irrespec ve of the specific work involved.
A project life cycle can vary from predetermined or plan-driven approaches to flexible or
change-driven approaches. In a predetermined project life cycle, all things are defined at the
star ng stage and any altera ons to scope are carefully addressed. In a flexible life cycle, the
product is developed over mul ple itera ons. During the itera ons, the func onali es of the
cycle will be discussed, implemented and, finally, reviewed by the client. This life cycle is
generally followed by the organiza ons to reduce risk of uncertain es in a project.
In general, the four phases of a project life cycle are as follows:

Initiation Phase Planning Phase Implementation Phase Execution Phase

As you know, Ar ficial Intelligence is a branch of Computer Science that is a study of how the
human brain thinks, learns, decides and works, when it tries to solve problems. Eventually, this
study outputs intelligent so ware systems. In order to implement these systems, all
organiza ons need to adopt a comprehensive approach to cover each step of the AI or machine
learning lifecycle. The steps involved in an AI project life cycle are as follows:
a) Problem Scoping: Problem Scoping is the first step of an AI project life cycle. In this step,
you have to write about the nature and boundaries of a problem. In simple words, it is a
process by which a problem is defined.
b) Data Acquisi on: Data Acquisi on or Data Gathering refers to a process of iden fying and
gathering all data required for an AI project.

PM Publishers Pvt. Ltd.

65
c) Data Explora on: Data Explora on refers to a process of
understanding the nature of the data you have to work 1)
with, in terms of quality, characteris cs, etc. This step Problem
provides you a be er understanding of data resul ng Scoping
in an effec ve end product.
d) Modelling: An appropriate analysis of collected 2)
5) Data
data is an essen al requirement for modelling Evaluation AI
Project Acquisition
data. In this step, you can select the datasets
Life
which match your requirements and train the Cycle
model using appropriate machine-learning
algorithms or AI-enabled algorithms.
4) 3)
e) Evalua on: Deployment is the last step of the AI Modelling Data
project life cycle. But, before deploying any AI- Exploration
enabled system, proper evalua on of a trained model is
necessary as it determines the accuracy of the model
according to the project requirements.

Let us understand more about the AI project life cycle with the help of a simple example.
Suppose you are going to organize a cultural fest of your school in a reputed hotel. Now, a list
of guests is prepared and invita ons are sent to them. According to your school management
team, no unauthorized person should be there. To ensure this, you can do the following things
to collect data:
a) Mandate entry in the hotel through an invita on card.
b) Mandate entry through a secret code which is given by you to all invitees.
c) Provide a separate ID card at the me of invita on.
The data collected by you is available in visual form. This data should be reliable to ensure the
smooth func oning of your school fest. In technical terms, the process of collec ng data is
known as Data Acquisi on.
A er collec ng data, now you would want to interpret some useful informa on of all invitees
like designa on, level, etc. To do this, you can explore the collected data and put it in an
organized way. The process of exploring data to interpret some useful informa on is known as
Data Explora on.
A er data explora on, now you would want to develop a system which detects the unique ID
of a person at the me of entry. This ID should be matched with the exis ng ID that you have
in your system. To do this, you can feed all the mandatory data in an AI intelligent system like
biometric and train it in such a manner that it recognizes an unknown person immediately. In
technical terms, the process of training a model using a machine learning algorithm based on
accurate data is called the Modelling stage.
A er training the model, you can deploy it and test it by pu ng accurate and inaccurate data.
This stage is known as the Evalua on stage.
DO YOU
KNOW ?
Data is the most founda onal element that
makes AI highly powerful.

Artificial Intelligence - 9
66
 ICE-BREAKER ACTIVITY
AI PROJECT LIFE CYCLE
You have learnt various stages of the AI project life cycle in which the data acquisi on stage
plays an important role to train a model. Now, explain the importance of data explora on in
your own words.

The detailed descrip on of AI project life cycle is as follows:

Problem Scoping
Problem Scoping is the ini al stage of the AI project life cycle. The term ‘Problem Scoping’
refers to a process of framing a problem in such a way that you will have a vision to solve it. In
general, it involves a series of steps to narrow down to a problem statement from a broad
theme. In simple words, it is basically selec ng a problem which we might want to solve using
our AI knowledge. In this session, we’ll learn how to scope a problem with the help of the
following worksheets:

Worksheet 1: Narrow Down a Problem Statement from a

Broad Theme

In your daily life, you must have seen various things around you like water, birds, animals,
etc. Select any theme according to your interests or from the suggested themes given below:

Sanitation
Women
Safety Environment

Cyber
Traffic
Themes Safety

Disability Health

Education

From the given themes, select any one theme or think of any other theme on your own. In
the previous chapter, you have learnt 17 goals of Sustainable Development which are also
based on themes. Therefore, you can also select any theme using the 17 Sustainable

PM Publishers Pvt. Ltd.

67
Development Goals. A er selec on, you can make a list of topics and find out various problems
which exist in a par cular theme. These problems will now be very specific as they have been
narrowed down from a broader sense.
Write the goal of your project a er selec ng any one problem that you have scoped to get a
clear vision of the problem that you are looking forward to solve using your AI knowledge. Now,
fill the following details on the basis of your interests:
Selected theme for your project: ……...............................................……………………………………………….
List of topics: ………...............................................…………………………………….………………………………………
………...............................................……………........................………………………………………………………………
Choose any one problem that you listed down under the topic sec on.
Now, you can set up the goal of your project by selec ng one problem out of the ones listed
above, which you want to solve using your AI knowledge. This problem now becomes the target
of your AI project and helps you to get a clear vision of what is to be achieved. Let us now
state the selected problem as a goal. For example, a goal can be stated as, “How might we help
teachers to determine the best method of inclusive educa on?”
Now, write the goal or aim of your project in the space given below:
………...............................................……………........................………………………………………………………………

Worksheet 2: 4Ws Problem Canvas

Consider the goal selected in Worksheet 1 as the problem of this ac vity. You can use 4Ws
framework to understand the procedural steps of problem solving. The 4Ws are the basic
ques ons used to collect informa on regarding a problem. So, what are the 4Ws? The 4Ws
refer to Who, What, Where and Why.

Who

4Ws
Why What
Framework

Where

Artificial Intelligence - 9
68
 The detailed descrip ons of the 4Ws are as follows:
a) Who: This block helps us in iden fying the stakeholders or the team that will work on
the problem. At this stage, we are looking at the person who is ul mately benefi ng
from the project comple on. Various types of ques ons under this block are: Who are
the stakeholders?, Find detailed informa on about the stakeholders, etc.
b) What: This block helps us in determining the nature of the problem. What is the
problem and how do you know that it is a problem? At this stage, a requirement to
gather evidence to prove that the problem you have selected actually exists, is
mandatory. You can gather evidence through various mediums like newspaper ar cles,
media, announcements, Internet, etc. Various types of ques ons under this block are:
What is the problem?, List at least two pieces of evidence which show that it is a
problem, etc.
c) Where: Under this block, you can focus on the situa on or the context of the problem.
This block will help you look into the situa on in which the problem arises, its context,
and the loca ons where it is notable. Various types of ques ons under this block are: At
which point or loca on do you think it is a problem?, What is the situa on in which the
stakeholders experience the problem?, etc.
d) Why: Under this block, you can think about the benefits which the stakeholders would
get from the project comple on and how it would benefit them as well as the society.
Various types of ques ons under this block are as follows: How would a stakeholder get
a benefit from the solu on?, Why will stakeholders prefer this solu on?, etc.
A er understanding in brief about the 4Ws, you should be able to summarize all the key
points of the 4Ws in a paragraph form. Use the following template if necessary:

Our .............................................................................. [brief descrip on of the stakeholders]

....................................................... has/have a problem that ...........................................
[issue, problem, need] ............................................ when/while ............................................
[context, situa on]. An ideal solu on would be ................................................. [benefit of
the solu on for them] .............................................................................. .

Data Acquisition
You have learnt that data is the core of any Ar ficial Intelligent system. Thus, data acquisi on
plays an important role in the AI project life cycle. As you know, Ar ficial Intelligent systems
have the ability to process enormous amounts of data and their accuracy increases along with
the data volume. Data can be a piece of informa on or facts and sta s cs collected together
for reference or analysis. We feed data into an AI model and train it to predict desired output.
Let us understand the concept of dataset with the help of a simple illustra on. Suppose you
want to develop an AI-enabled system which can predict the presence or absence of any
employee on the basis of facial recogni on. To do this, you would feed the data of his/her face
into the machine. This is the data with which the machine can be trained. Now, once it is ready,
it will predict his/her a endance in an effec ve manner. The facial characteris cs of the

PM Publishers Pvt. Ltd.

69