National Rail and Transport Institute

(Deemed to be University)
National Academy of Indian Railways Campus,
Lalbagh, Vadodara, Gujarat - 390004
www.nrti.edu.in | 0265-2648305

Internship Report

AI & ML IN MODELING, PREDICTTON AND MAINTAINANCE

OF THE SYSTEM

A report submitted
In partial fulfilment of the requirements
for the degree of
B. Tech/ B. Ba/ B. Sc
By

NRTI Internship Group- DRM Mumbai Central/BCT-WR

To
DRM Mumbai Central/BCT-WR
&
National Rail and Transportation Institute (NRTI)
Vadodara, India-390004
Date: August 30, 2021
 National Rail and Transport Institute
(Deemed to be University)
National Academy of Indian Railways Campus,
Lalbagh, Vadodara, Gujarat - 390004
www.nrti.edu.in | 0265-2648305

SUMMER INTERNSHIP
Certificate of Completion

This is to certify that the work contained in the project titled “AI &ML in
Modelling, Prediction and Maintenance of the System” has successfully
completed by the following students under my/our supervision during the 4-week
summer internship at Mumbai Central/BCT-WR. During the internship period the
students found to be hard working, punctual and discipline and they met all the
criteria satisfactorily.

Name Course Roll No

Abhishek Chandra Bhagwat B. Tech 203005
Parakh Srivastava B. Tech 203058
Pradumn Kumar B. Tech 203060
Subrat Dash B. Tech 203093
Prasad Prakash Kedare B. SC 201044
Rohan Manoj Sharma B. SC 201055
Harsh Vardhan Rathore B. SC 201027
Jyoti Prakash Mishra B. SC 201032
Mitali Pedhekar B. Ba 202024
Prathmesh Badgujar B. Ba 202026
Tejas Manoj Patil B. Ba 202042
Ayush Ranjan Srivastava B. Ba 202015

Signature
Date: August 30, 2021 Industry Mentor details
 National Rail and Transport Institute
(Deemed to be University)
National Academy of Indian Railways Campus,
Lalbagh, Vadodara, Gujarat - 390004
www.nrti.edu.in | 0265-2648305
Acknowledgement

We take this opportunity to express our sincere thanks to Shri G.V.L.

Satya Kumar, DRM/MMCT/WR for having us this opportunity to frame
our report on “AI & ML IN MODELING, PREDICTTON AND MAINTAINANCE OF THE
SYSTEM” and our respected mentors Poonam Kumari Saha, Assistant
Mechanical Engineer (Mumbai suburban),Krutika Raut (Junior
Engineer-Signal & Telecom), for their guidance and support. It was a
privilege to have a great experience working under them in a pleasant and
friendly environment. We are thankful to office staffs and other relevant
persons.

We are also very much thankful to our respective professors Dr. Hari
Krishna Gaddam, Dr. V. Chintala, Dr. Sunil Kumar Sharma, Dr.
Abhilasha Saksena for their guidance and support.

We are very thankful to the National Rail and Transportation

Institute (NRTI) and Indian Railways, for providing us with the
opportunity of pursuing summer internship in Mumbai Central [Western
Railways].

We are also thankful to Shri Hublal Jagan, Registrar, NRTI and

Shri Krishan Kumar, Dy. Registrar, NRTI, Jayaraju Poreddy,
Research and Teaching Associate, NRTI for believing in us.

In the end we would like to express our gratitude towards our parents
for their kind cooperation and encouragement which helped us in completion
of this project.
 National Rail and Transport Institute
(Deemed to be University)
National Academy of Indian Railways Campus,
Lalbagh, Vadodara, Gujarat - 390004
www.nrti.edu.in | 0265-2648305

Abstract
In this internship we went through the different aspects of AI. Artificial
intelligence focuses on development of machines which are programmed to think
more like a human and use this learning to make predictions about future states. AI
plays a significant role in various fields. Mostly we are talking about the advantages
of AI, but far behind to discuss its limitation too. We are deprived from the main goal
of AI which is AGI rather we are only focusing in the field of ANI. ANI had developed
in major sectors which had a positive aspect.

Using AI whether in our household or in the company is not as easy step. It involves
a series of steps such as selecting a favourable project, workflow of making a machine
learning project, data science project etc. This doesn’t require 100 members to work
upon. It can be started by a single engineer, or by ourselves by referring through
various online courses and get some idea. AI also plays a major role in society. People
presumes AI in different manners. Some becomes too optimistic and other too
pessimist. The sets of nos. Used becomes bias.

As it widens the gap between women and men, rich and poor. Many spammers steal
information by using illegal means through which data are stolen and several crimes
are committed. Researchers are working to eliminate these limitations. Research
shows that around 400 to 800 million will be displaced due to automation but it is
also likely to create more jobs as more skilled human resource will be requires with
expansion of AI. AI provides scope for more development in the developing
economies. But it also be kept I mind that AI needs a lot of work to be done, so these
nations should make education the first priority. For AI to work better more training
and data is required.

Eventually, we had discussed major points of AI and gone through its crux. We
looked at various benefits and limitation and provided recommendations about how
it should be tackled so that the mentors can have a look and take necessary steps and
changes in the overall aspect.
 National Rail and Transport Institute
(Deemed to be University)
National Academy of Indian Railways Campus,
Lalbagh, Vadodara, Gujarat - 390004
www.nrti.edu.in | 0265-2648305

Table of Contents

Acknowledgement.................................................................................................................. 3
Abstract.................................................................................................................................. 4
Chapter 1: Introduction........................................................................................................... 7
1.1 Motivation................................................................................................................. 7
1.2 Objectives................................................................................................................ 7
1.3 Organization of Report.............................................................................................7
Chapter 2: Exploratory Data Analysis.....................................................................................8
2.1 MS Excel....................................................................................................................... 8
2.2 Google Data Studio.......................................................................................................9
Chapter 3: Artificial Intelligence & Machine Learning............................................................10
3.1 Data............................................................................................................................ 10
3.2 Machine learning.........................................................................................................11
3.3 Artificial Intelligence....................................................................................................12
Chapter 4: AI vs ML.............................................................................................................. 14
4.1 Expectations VS Reality..............................................................................................14
Chapter 5: Building A.I. Projects...........................................................................................15
5.1 STEPS........................................................................................................................ 15
5.2 Workflow of making a Machine Learning project.........................................................15
5.2.1 Working steps of a speech recognition system........................................................15
5.2.2 Working steps of a self-driving car........................................................................16
5.3 Workflow of making a Data Science project................................................................16
5.3.1Choosing an A.I. Project to work on......................................................................17
Chapter 6: Building A.I. in a company / organization............................................................18
6.1 Role of an AI team......................................................................................................18
6.2 AI Transformation Playbook........................................................................................19
6.2.1 Case Study on Complex AI product......................................................................20
6.3 Portfolio of AI.............................................................................................................. 21
Chapter 7: A.I. and Society...................................................................................................22
7.1 A realistic view of AI....................................................................................................22
 National Rail and Transport Institute
(Deemed to be University)
National Academy of Indian Railways Campus,
Lalbagh, Vadodara, Gujarat - 390004
www.nrti.edu.in | 0265-2648305
7.2 Discrimination / Bias...................................................................................................22
7.3 Adversarial attacks on AI............................................................................................22
7.4 AI and developing economies.....................................................................................23
7.5 AI and jobs.................................................................................................................. 23
Chapter 8: TRAIN MANAGEMENT SYSTEM (TMS)............................................................24
8.1 Advantages of setting up TMS in Mumbai suburban...................................................25
A Run-Through AI................................................................................................................25
FUTURE OF AI................................................................................................................. 26
Overview of Internship Activities...........................................................................................27
REFERENCES:.................................................................................................................... 29

List of Figures
Figure 1 Pivot Table.........................................................................................................................................................8
Figure 2 Google Data Studio..........................................................................................................................................9
Figure 3 AI in Business Units......................................................................................................................................19
Figure 4 AI Implementation Cycle............................................................................................................................19
Figure 5 Self-Driving Car Working System............................................................................................................20
Figure 6 First TMS that was inaugurated in Western Railways to improve Mumbai local train
operations...............................................................................................................................................................24
 National Rail and Transport Institute
(Deemed to be University)
National Academy of Indian Railways Campus,
Lalbagh, Vadodara, Gujarat - 390004
www.nrti.edu.in | 0265-2648305

Chapter 1: Introduction
1.1 Motivation
The Indian Railways keep a record of everything in the computers to analyse
different fields of things in detail. After analysis, if problems occur, steps can be
taken to prevent the problems. Officers and Managers get the best idea of the past
events when they have the record in front of them. Usually, data is always in table
format in either Microsoft Excel or Google Sheets. This data can be turned into a
much more informative manner by using tools like “Pivot Charts” in Microsoft Excel
and “Google Data Studio” of Google. In this manner, the officers can understand the
occurred events, identify mistakes and come out with different and effective
solutions for the profit and betterment of Indian Railways.

1.2 Objectives
The main objectives of this internship/study are-
 To know about artificial intelligence and machine learning and how it is used
in modelling, prediction and maintenance of system.
 Familiarize ourselves that how google data studio and pivot tables are used in
analyzing the types of complains, framing possible solutions for the same.

Side objectives of this internship/study are-

 To know about Train Management System (TMS), and how it is used in
controlling and regulating the suburban traffic.
 To know that how image sensor camera is used in prediction of fault in Brake
block system and bogie.

1.3 Organization of Report

This report is organized considering data as the building blocks. Initially in the Parts
1 and 2 of the report data is filtered and then structured in graphs using MS Excel's
pivot table and the similarly charts are used to create dashboards in Google Data
Studio which using Google Sheets as data source. These graphs and charts help us to
visualize the overall pattern of the data and hence, is very useful.
Further in Part 4 onwards, we see how data is valuable in Artificial Intelligence
(more importantly, accurate data which is usable for the AI’s development purpose).
 National Rail and Transport Institute
(Deemed to be University)
National Academy of Indian Railways Campus,
Lalbagh, Vadodara, Gujarat - 390004
www.nrti.edu.in | 0265-2648305

Chapter 2: Exploratory Data Analysis

2.1 MS Excel
To make a Pivot Chart, the first step is to clean the received data from the
different offices from the respective division. The data is recorded in different forms.
It has to be rearranged in a specific form before using the pivot chart. The data is
then converted into a table and then pivot charts are made out of the data table.
These Charts are much more informative than the raw data from the data table.

Figure 1 Pivot Table

 National Rail and Transport Institute
(Deemed to be University)
National Academy of Indian Railways Campus,
Lalbagh, Vadodara, Gujarat - 390004
www.nrti.edu.in | 0265-2648305

2.2 Google Data Studio

Using Google Data Studio is the upgraded and much more detailed version of
Pivot Charts which is available in Microsoft Excel. There are better graphics and
more designing options to make your data representable. New and different types of
charts options are available in Google Data Studio. The first step to use Data Studio is
to clean and import your data from Microsoft Excel to Google Sheets. After uploading
the data in Sheets, insert the Data Sheet in the Data Studio. The Data Studio will
automatically detect Headers and the content in it. There are many options to
customise your chart as per needed. The information is displayed quickly and in a
detailed manner.

Figure 2 Google Data Studio

 National Rail and Transport Institute
(Deemed to be University)
National Academy of Indian Railways Campus,
Lalbagh, Vadodara, Gujarat - 390004
www.nrti.edu.in | 0265-2648305

Chapter 3: Artificial Intelligence & Machine

Learning

3.1 Data

Data is messy. Big data or huge chunks of data, used properly, are really useful for a
business to prosper. It helps the machines to learn a certain task using ML and AI
algorithms. To achieve a proper dataset for ML and AI, first, you have to carefully
divide the input and the desired output part so as to make it useful.

Let us take an example of data that an agency might have, when they tried to
price the houses according to its size and number of bedrooms. So, the size and
number of bedrooms are the input (A) and the price is the output (B).

Suppose we have a fixed budget and we want to know the size of house we can afford.
So, this difference in purpose totally changes the input and output. Here, the price is
the input (A) and the rest left is the output (B).
 National Rail and Transport Institute
(Deemed to be University)
National Academy of Indian Railways Campus,
Lalbagh, Vadodara, Gujarat - 390004
www.nrti.edu.in | 0265-2648305

There are many ways of collecting this data for ML and AI.

1) Manual Labelling: - We can collect a dataset where a set of images are the input
and labels are the output. It is also known as a labelled data.

2) Observation: - We can even observe the behaviour and conditions of the object as
input and predict the required output.

3) Using Available data: - We can download it from a website or get it from a partner.

Only collecting data will not be beneficial, we have to be careful what type of data to
collect, so as to achieve our goal or output. How to collect the data so it can be used
in the right direction. So, data engineering is as important as data collection.

3.2 Machine learning

Machine Learning is a subset of Artificial intelligence. It is the field of study that

gives computers the ability to learn without being explicitly programmed. An
algorithm that learns to map from input to output, a.k.a., Supervised Learning, is the
most common type of Machine Learning.
Some of the examples and its useful applications, you can relate to are: -
 National Rail and Transport Institute
(Deemed to be University)
National Academy of Indian Railways Campus,
Lalbagh, Vadodara, Gujarat - 390004
www.nrti.edu.in | 0265-2648305
We are capable enough to collect huge amounts of useful data and more data implies
better performance of the Machine Learning system.But, keeping an eye on the facts,
Supervised Learning or Machine Learning is not taking off the way it's expected. The
below mentioned graph can justify our facts.

More data certainly means better performance. But after a certain point, it doesn't
get much better. However, neural networks and Deep Learning have made the
situation better, as it shows in the graph below: -

3.3 Artificial Intelligence

Artificial Intelligence is like the simulation of human intelligence by machines.

Much like I'm ML, AI takes huge chunks of labelled data to search for correlations
and patterns, and use this learning or intelligence to make predictions about future
states, like the responses given by a chatbot.
However, this has had a great impact almost in every aspect of our life and in every
field of study. It also has high expectations for the future. The statistics below will
prove the point: -
 National Rail and Transport Institute
(Deemed to be University)
National Academy of Indian Railways Campus,
Lalbagh, Vadodara, Gujarat - 390004
www.nrti.edu.in | 0265-2648305

There are two types of A.I:

 National Rail and Transport Institute
(Deemed to be University)
National Academy of Indian Railways Campus,
Lalbagh, Vadodara, Gujarat - 390004
www.nrti.edu.in | 0265-2648305

Chapter 4: AI vs ML

4.1 Expectations VS Reality

We all know how AI is making a huge difference in all the fields. Hence, we
have kept huge expectations from AI. Our expectations have become a little out -
reached. There are even discussions about AI robots conquering the Earth, and
unexpected imaginations about AI developments in the near future. But the truth is
everyone is speaking about the positive and achievements in AI. No one is talking
about the failures; the bigger projects AI couldn't achieve. AI robots are a whole
different reality.

The development we see is only in the field of AI is only in the field of ANI (Artificial
Narrow Intelligence) like smart speakers, chatbots, and self-driving cars. But there is
even a second part of AI, i.e., AGI (Artificial General Intelligence). Actually, that was
the main goal of building AI, but unfortunately, there is almost no progress or zero
development in this field.

The major development in ANI and the popularizing of the positive results of AI has
made people expect more and imagine so much as some clever robot, coming to take
over humanity. But we have to get out and face the reality, keep on trying and
researching to develop this amazing field of AI. We must also know the capabilities
and limitations of AI. There are several examples of businesses and companies,
investing blindly in huge amounts in AI, yielding no results. So, proper research and
realisation of the reality is needed, before taking any such step.
 National Rail and Transport Institute
(Deemed to be University)
National Academy of Indian Railways Campus,
Lalbagh, Vadodara, Gujarat - 390004
www.nrti.edu.in | 0265-2648305

Chapter 5: Building A.I. Projects

5.1 STEPS
To use AI in our household works or at a company we have to go through a series of
steps to do an AI project.

The steps are –

• Workflow of the AI project.
• Selecting a favourable AI project.
• Organizing the data at hand and building a team for the project.

5.2 Workflow of making a Machine Learning project

Algorithms of machine learning are trained by input to output or in simple
words, mapping from A to B.
The working of machine learning can be understood by a speech recognition system.

5.2.1 Working steps of a speech recognition system

1. A lot of audio data is collected by recording audio.

2. The model is trained by a machine learning algorithm so that system begins to

recognize data that is provided to it. This process is repeated many times to train
the model over and over again.

3. The trained AI software is put in a smart speaker and it is ready for shipping.
 National Rail and Transport Institute
(Deemed to be University)
National Academy of Indian Railways Campus,
Lalbagh, Vadodara, Gujarat - 390004
www.nrti.edu.in | 0265-2648305

The smart speaker may not work fully efficiently from the very beginning due to
being exposed to a lot of real-world data which involves coming across a lot of
different accents. Nonetheless, the model receives a lot of data which helps in
updating the quality of the model.

5.2.2 Working steps of a self-driving car

1. A lot of data is collected in the form of images and various positions of

different cars.

2. The model is trained by a machine learning algorithm so that it recognizes

images of cars in various positions. This process is repeated many times.

3. The model is deployed to get real-world data which shows the shortcomings of
the model like being unable to identify a less seen vehicle like a golf cart.
Information like this is crucial in updating the model.

5.3 Workflow of making a Data Science project

Following steps constitute the making of a data science project –

1. Collecting data. For example, collecting data of the customers that are placing
an order on a website.

2. Analysing data. For example, analysing customer’s shopping habits and

improving accordingly can very well help increase the sales of an online shopping
website. These analyses are repeated many times to get valuable insights.

3. After deploying new measures which are brought into practice by analysing
data, new data are re-analysed from time to time to find if even better insights are
available.
 National Rail and Transport Institute
(Deemed to be University)
National Academy of Indian Railways Campus,
Lalbagh, Vadodara, Gujarat - 390004
www.nrti.edu.in | 0265-2648305

Tools like data science and machine learning can help any job function with the
availability of data like in the field of sales, recruiting, marketing, manufacturing,
agriculture, etc.

5.3.1Choosing an A.I. Project to work on

A team consisting of AI experts along with Domain experts brainstorm their

ideas so that they can select an AI project to work keeping in mind that the work can
be done by an AI and it is useful for the business the work is being done for.

Steps to build such a framework -

• The goal should be to look for individual tasks that can be automated instead of
automating jobs as a whole.

• Finding the main driving force of business values for said business.

• Finding main recurring problems in the business.

 National Rail and Transport Institute
(Deemed to be University)
National Academy of Indian Railways Campus,
Lalbagh, Vadodara, Gujarat - 390004
www.nrti.edu.in | 0265-2648305

Chapter 6: Building A.I. in a company /

organization

AI products may need large AI teams. But it’s always good to gain some momentum
starting with small teams. We don't require 100 people for AI projects. And we can
even get started with one Software Engineer, a single Machine Learning Engineer, a
single Data Scientists, or may even nobody, but ourselves, by taking some online
courses on machine learning or deep learning or data science, that's often enough to
start looking at some smaller volumes of data, start drawing some conclusions or
start trading some machine learning models to get going.

6.1 Role of an AI team

1. Software Engineer: To write specialized software for execution of the task.

2. Machine learning Engineer: To generate input-output mapping and to build

another Machine Learning algorithm.

3. Machine Learning Reacher: To extend the state of art in ML

4. Data Scientists: To examine data and provide insights and present to team and
executives.

5. Data Engineer: Organize data and stores in an easily accessible and cost-effective
way.

6. AI product Manager: Decides what to build, what is feasible and valuable.

 National Rail and Transport Institute
(Deemed to be University)
National Academy of Indian Railways Campus,
Lalbagh, Vadodara, Gujarat - 390004
www.nrti.edu.in | 0265-2648305

Figure 3 AI in Business Units

6.2 AI Transformation Playbook
1. Execute trail projects to gain Momentum

2. Build an insource AI team

3. Provide AI training to employees

4. Make an AI strategy

5. Develop internal and external communication

Figure 4 AI Implementation Cycle

 National Rail and Transport Institute
(Deemed to be University)
National Academy of Indian Railways Campus,
Lalbagh, Vadodara, Gujarat - 390004
www.nrti.edu.in | 0265-2648305
6.2.1 Case Study on Complex AI product

E.g.- Self Driving Car

Figure 5 Self-Driving Car Working System

 National Rail and Transport Institute
(Deemed to be University)
National Academy of Indian Railways Campus,
Lalbagh, Vadodara, Gujarat - 390004
www.nrti.edu.in | 0265-2648305
6.3 Portfolio of AI
 National Rail and Transport Institute
(Deemed to be University)
National Academy of Indian Railways Campus,
Lalbagh, Vadodara, Gujarat - 390004
www.nrti.edu.in | 0265-2648305

Chapter 7: A.I. and Society

7.1 A realistic view of AI

AI plays an important role in society. AI is presumed differently by different
peoples like the Goldilocks story. Some of them become too optimistic while others
become too pessimistic. It should be kept in mind that AI has both limitations as well
as advantages. AI has been a boon in many organizations but also failed in several
areas such as explain ability.

7.2 Discrimination / Bias

AI exhibits discrimination in various forms. AI compares men as more
superior to women, light skin tone as smarter than dark skin tone. This widens the
gap between two groups. It happens because AI uses different sets of numbers. Using
0 in the place of that number will remove that bias factor.

7.3 Adversarial attacks on AI

AI can be fooled by making minor changes which in reality are not any
significant change to be thought upon, this can be said as an attack on AI. Many
online frauds take place such as money is stolen from accounts etc. Researchers are
working to put a strong defence system against such attacks but it involves some cost.

Adverse uses of AI. AI has done numerous things for the welfare of people but
besides this it has also some adverse uses. It can make things appear in a way which
one has not done ever? There are spams and anti-spams battles, fraud and anti-fraud
battles. To eliminate this, there are a lot of resources available on the anti-spam and
anti-fraud side, on which AI has to work a lot.
 National Rail and Transport Institute
(Deemed to be University)
National Academy of Indian Railways Campus,
Lalbagh, Vadodara, Gujarat - 390004
www.nrti.edu.in | 0265-2648305

7.4 AI and developing economies

AI helps to enhance nations wealth. In developing nations, it provides new

ways to increase productivity but less opportunities for the lower wage groups. AI
lags and somehow proves immature in developing economies. There should be more
spending on education because there is a pool of this which is to be learnt in AI
before completely utilizing in developing economies.

7.5 AI and jobs

As per the research 400 to 800 million will be displaced by 2030 due to
automation. But on the other side the same research shows that the jobs created by
AI will be even larger. The jobs that include repetitive work are more likely to be
automated and the work which requires more social interaction is less likely to be
automated.
 National Rail and Transport Institute
(Deemed to be University)
National Academy of Indian Railways Campus,
Lalbagh, Vadodara, Gujarat - 390004
www.nrti.edu.in | 0265-2648305

Chapter 8: TRAIN MANAGEMENT SYSTEM

(TMS)
Train management system in India was established in Mumbai Central [it records
train movements from Churchgate to Virar] a division of Western Railways to get
real time data of –
 TRAIN MOVEMENTS

 STATUS OF SIGNALS
 BLOCK SECTIONS
 INTERLOCKING SYSTEM
 LEVEL CROSSING GATES

The system is based on latest technology of digital light processing projector and
it tracks approximately 1049 trains on this route. It was launched in 2004. This
system is connected to various departments in that particular section of western
railways.

TMS can also regulate diversion of trains, induction of trains and reversal of trains
in the territory. It also allows the controllers to set up the routes for the trains and
prioritize handling of trains.

Figure 6 First TMS that was inaugurated in Western Railways to improve Mumbai local train operations
 National Rail and Transport Institute
(Deemed to be University)
National Academy of Indian Railways Campus,
Lalbagh, Vadodara, Gujarat - 390004
www.nrti.edu.in | 0265-2648305

8.1 Advantages of setting up TMS in Mumbai suburban

 It increased the line capacity by reducing the gap between two trains, it used
the real time availability of train movements.
 It improved the punctuality of local trains.
 Faster response to emergencies can be provided.
 On time accurate and uniform availability of real time data to all traffic
operators resulting in seamless operations during taking over of trains.
 It removed the older system of listing down the information of trains and
frequently calling station master for that.
 If the TMS can be started in every part of the Indian Railways (busy sections)
then Indian Railways can save up to millions every year by cutting down the
number of employees.

A Run-Through AI

Since, the first Turing test, we have come a long way. There is still a much more to
be done. Out of it’s three stages ANI also known as Weak AI, AGI also known as
strong AI and the third speculated, ASI (Artificial super intelligence), we have marked
some achievement only in ANI. The types of AI include - Relative Machine AI
(operate solely on present data), Limited memory AI (makes informed and improved
decision based on past memory), Theory of mind AI (Mainly focuses on emotional
intelligence), Self-Aware AI (with its own consciousness).

At present, most of the forward-thinking business/enterprises are working full time in

development/improving of NLP (natural language processing), it is one of the
domains of the AI. NLP bridges the gap between human and machines. It also
includes automating of call centres. Customer experience is huge driver for a
company’s profit. Earlier, a customer with a query had to wait a long time in queue to
get addressed. Today, NLP can help with the basic problems and at least it transfers
the call to the right person, saving all that ‘going through different representative’.
 National Rail and Transport Institute
(Deemed to be University)
National Academy of Indian Railways Campus,
Lalbagh, Vadodara, Gujarat - 390004
www.nrti.edu.in | 0265-2648305
Ethical challenge. Some of the ethical challenges are Biases (it is important to feed
diversified data to the algorithm), Control and Morality of AI, privacy, power balance,
environmental impact.

FUTURE OF AI

For AI to work better it needs to be trained better. For its better training, it needs
to be fed large amount of data and for it to get processed it needs next level
processing capabilities. Earlier it was only based on the processor, then came the
Graphic Cards- it has been proven very useful in improving efficiency. Quantum
computing along with other next level processing resources such as biological and
neuromorphic computing will unlock more capabilities for AI.

Quantum computing is theoretically capable of 100 trillion times faster than

today’s fastest computer. One of the most sophisticated generative AI models
already contains over 175 billion different parameters. With quantum computing,
machine learning model will become inevitable larger which in result will work
efficient on more parameters. Today the most difficult part is to collect more real
data, but the high processing power we will be able to create larger “synthetic data”.
The next can be creative AI, it also depends on the development of generative AI. It
includes development of art, music, poetry.

Today some creative AI can also write computer code, although it is in its early
phase. For recent contribution of AI and Importance, we can take the example of-
How AI contributed in R&D of COVID-19 vaccines. AI and Machine learning were
able to sort through the huge data stored in outer protein, and predict the most
immunogenic protein- which guided researchers in producing targeted vaccine.
 National Rail and Transport Institute
(Deemed to be University)
National Academy of Indian Railways Campus,
Lalbagh, Vadodara, Gujarat - 390004
www.nrti.edu.in | 0265-2648305

Overview of Internship Activities

Day 1 – August 2, 2021, 10 AM to 2 PM

1. Data Analysis using Pivot Table using data collected from complaints of passengers from
Railway's website.
2. Introduction to Google Data Studio creating charts from the data provided and importing images.

Day 2 – August 3, 2021, 10 AM to 2 PM

1. Analysis of real time data using pivot table, visualizing the pivot table using charts.
2. Importing data into Google data studio from source and getting started.

Day 3 – August 4, 2021, 10 AM to 2 PM

1. Solving problems using pivot table – creating charts and graphs using pivot table.
2. Adding data source for Google Data Studio by loading Google Sheets. Creating dashboards from
previously created charts and imported images to represent the data acquired from Railway’s
website.

Day 4 – August 5, 2021, 10 AM to 1:30 PM

Basics of AI – Classification of AI into Artificial Narrow Intelligence (ANI) and Artificial General
Intelligence (AGI), their differences and scopes.

Day 5 – August 6, 2021, 10 AM to 2 PM

Machine Learning and its basics with the concept of supervised learning.

Day 6 – August 9, 2021, 10 AM to 2 PM

Data – its definition and the need to collect the correct type of data for the required purpose.
Acquiring data by manual entry, from observing behaviour, download from the internet.

Day 7, August 10, 2021, 10 AM to 2 PM

Need for working alongside AI team before collecting data - so that they might not end up collect
unnecessary data. The feedback from the AI team is also very important.
Being mindful of incorrect labels, missing values etc. In data

Day 8, August 11,2021, 10 AM to 2 PM

Classification of multiple types of data.
Dealing with unstructured data.

Day 9, August 12,2021, 10 AM to 2 PM

Understanding the terminology – Machine Learning, Data Science, Neural Network, Deep Learning

Day 10, August 13, 2021, 10 AM to 2 PM

The working of Neural Network

Day 11, August 16, 2021, 10 AM to 2 PM

Stepwise guidelines for making an AI company
 National Rail and Transport Institute
(Deemed to be University)
National Academy of Indian Railways Campus,
Lalbagh, Vadodara, Gujarat - 390004
www.nrti.edu.in | 0265-2648305
Day 12, August 17, 2021, 10 AM to 2 PM
The abilities and limitations of Machine Learning and how Machine Learning eases problem solving.

Day 13, August 18, 2021, 10 AM to 2 PM

Differences between Deep Learning and Neural Network

Day 14, August 19, 2021, 10 AM to 2 PM

Introduction to Face Recognition system
Feeding data using Face Recognition system.

Day 15, August 20, 2021, 10 AM to 2 PM

Workflow of AI projects
Workflow of Data Science projects

Day 16, August 21, 2021, 10 AM to 2 PM

Different roles in an AI team
Case study of complex AI projects

Day 17, August 22, 2021, 10 AM to 2 PM

How self-driving car AI takes inputs and ultimately decisions....
AI transformation playbook

Day 18, August 23, 2021, 10 AM to 2 PM

Executing initial projects to gain momentum

Day 19, August 24, 2021, 10 AM to 2 PM

Development of an AI strategy

Day 20, August 25, 2021, 10 AM to 2 PM

Dos and Don’ts for AI

Day 21, August 26, 2021, 10 AM to 2 PM

Improving AI to eradicate bias and adversarial attacks
 National Rail and Transport Institute
(Deemed to be University)
National Academy of Indian Railways Campus,
Lalbagh, Vadodara, Gujarat - 390004
www.nrti.edu.in | 0265-2648305

REFERENCES:
 AI for Everyone. Author: Andrew Ng
https://www.googleadservices.com/pagead/aclk?
sa=L&ai=DChcSEwj6tKHSldbyAhWQBXIKHVGaBIUYABAAGgJzZg&ae=2&ohost=w
ww.google.com&cid=CAESQOD2QqzZBLaTPsQ75Lzf2xmsm3HvONpPBhlHJsfm4H
41bcnarKFO-QHIOZR4vUfukd2TW-
TVjknrFjGjt7lzctY&sig=AOD64_04xsivGQAWg5KSa-
uaoJzV1e5n1w&q&nis=1&adurl&ved=2ahUKEwi39pjSldbyAhWDUn0KHZxVAU4Q0
Qx6BAgCEAE

 What is deep learning & how does it works?

https://searchenterpriseai.techtarget.com/definition/deep-learning-deep-neural-
network

 The impasse of Artificial Intelligence and cognitive science

https://www.researchgate.net/publication/228396927_The_Impasse_of_Artificial_Inte
lligence_and_Cognitive_Science

 What is definition of Machine learning?

https://www.expert.ai/blog/machine-learning-definition/

 How any Why business collect your personal data.

https://www.businessnewsdaily.com/10625-businesses-collecting-data.html