Yuvraj Vikrant Bharadia

Clevered AI
Expert Level
1
 Introduction
 My name is Yuvraj Vikrant Bharadia. I am
14 years old. I was born on 11th May 2009
and I am currently studying in Year 10C in
Dubai International Academy – Emirates
Hills. I have just finished my Clevered AI
Expert Level 1 course and am graduating
to AI developer level now. In this
presentation, I will give a comprehensive
overview of my Level 1 course. This will
include a brief introduction to Data
Analysis(Types of data analysis), Machine
Learning, Deep Learning and All related
syntax that I have acquired throughout
my journey with Clevered. At the end I
will be demonstrating the final project
which was a text summarisation project
which utilises neural networks to
summarise a long text by extracting key
words and phrases, similar to what we
have done before.
Acknowledgements

Thank you, Parents, Thank you, Teacher,

 I am deeply grateful to my  Heartfelt thanks to my dedicated
parents for their unwavering teacher, Mrs. Smita Bansal for your
support and financial exceptional guidance throughout
commitment to my AI course the AI course. Your meticulous
notes on challenging concepts and
since the beginning. Their belief
prompt timely responses outside
in my potential and sacrifice has
regular hours have been invaluable.
allowed me to focus on learning Your passion for teaching has not
without worrying about the only enhanced my understanding
burden of tuition. Thank you for but also inspired a deeper dive into
being my pillars of strength this path. Thank you for making my
morally and financially and could learning experience truly enriching
now have done it without you. and enjoyable.
 Data Analytics,
ML and AI
Types of Data Analysis:
 Descriptive: Load, Clean and Summarize data to identify patterns
and or meaning from it.
Methods: Data aggression and Data mining
Outcome: A visual representation of the data ie. Bar graph, Pie
chart and Line graph etc.
 Diagnostic: The technique is called asking our data “Why did this
happen?”
Methods: Analysing the graph to understand the root cause and
is used in the final project
Outcome: Seaborn sns heatmap to understand the most
correlating IVs affecting the DV.
 Predictive: Takes data from past events and company data and
analyses it to make future predictions through patterns and
trends
Methods: Linear regression, Decision Trees, Random Forest
classifiers.
Outcomes: Clear statistical data showing a prediction with a
certain probability
 Prescriptive: Recommendations for the future using all other
three aspects for example self-driving cars which make millions
of decisions per second.

Machine learning: Used to determine the likelihood of future events

Neural Network
A neural network is a method in artificial
intelligence (AI) that teaches computers
to process data in a way that is inspired
by the human brain. It is a type of
machine learning (ML) process, called
deep learning, that uses interconnected
nodes or neurons in a layered structure
that resembles the human brain.
1. Multiply all the inputs by their weights
w, real numbers that express how
important the corresponding inputs
are to the output,
2. Add them together referred as
weighted sum: ∑ wj xj,
3. Apply the activation function, in other
words, determine whether the
weighted sum is greater than a
threshold value, where -threshold is
equivalent to bias, and assign 1 or
less and assign 0 as an output.
 Artificial Neural Network
 The inputs on the left side represent the
incoming signals to the main neuron in the
middle. In a human neuron, this would include
smell or touch.
 In your Neural Network these inputs are
independent variables. They travel down the
synapses, go through the big grey circle, then
emerge the other side as output values.
 The main difference between the biological
process and its artificial counterpart is the level
of control you exert over the input values; the
independent variables on the left-hand side.
 But you can determine what variables will enter
your Artificial Neural Network.
 Now, once these values pass through the main
neuron and break on through to the other side,
they become output values.
MultiLayer Perceptron (MLP)
MLP:

They are comprised of an input layer,

a hidden layer or layers, and an output
layer. While these neural networks are
also commonly referred to as MLPs, it’s
important to note that they are
comprised of sigmoid neurons, not
perceptrons, as most real-world
problems are nonlinear. Data usually is
fed into these models to train them,
and they are the foundation for
computer vision, natural language
processing, and other neural networks.
Convolutional Neural
Networks(CNN)
CNN:
They are similar to
feedforward networks, but
they’re usually utilized for
image recognition, pattern
recognition, and/or
computer vision. These
networks harness
principles from linear
algebra, particularly
matrix multiplication, to
identify patterns within an
image.
Recurrent
Neural
Networks(RNN)
RNN:
They are identified by their feedback
loops. These learning algorithms are
primarily leveraged when using time-
series data to make predictions about
future outcomes, such as stock market
predictions or sales forecasting.
 Hence, Recurrent means the
output at the current time step
becomes the input to the next
time step. At each element of the
sequence, the model considers
not just the current input, but
what it remembers about the
preceding elements.
 This memory allows the network
to learn long-term dependencies
in a sequence which means it can
take the entire context into
account when making a
prediction.
LSTM(Long Short-term memory Neural Network
We calculate the gradient, that is, loss with respect to a particular set of
weights, adjust the weights accordingly and repeat this process until we
get an optimal set of weights for which loss is minimum.This is the
concept of backtracking.
Sometimes, it so happens that the gradient is almost negligible. It must
be noted that the gradient of a layer depends on certain components in
the successive layers. If some of these components are small (less than
1), the result obtained, which is the gradient, will be even smaller. This is
known as the scaling effect.
When this gradient is multiplied with the learning rate which is in itself a
small value ranging between 0.1-0.001, it results in a smaller value. As a
consequence, the alteration in weights is quite small, producing almost
the same output as before.
Similarly, if the gradients are quite large in value due to the large values
of components, the weights get updated to a value beyond the optimal
value.
This is known as the problem of exploding gradients.
They can learn long-term dependencies between time steps of data.

1.Input gate- It discover which value 3.Output gate- The input and the
2.Forget gate - It discover the
from input should be used to modify memory of the block are used to
details to be discarded from the
the memory. decide the output.
block.
Sigmoid function (σ) decides which Sigmoid function decides which
A sigmoid function decides it.
values to let through 0 or 1. values to let through 0 or 1 and tanh
It looks at the previous state (ht-1)
And tanh function gives weightage to function gives weightage to the
and the content input (Xt) and
the values which are passed, deciding values which are passed, deciding
outputs a number between 0(omit
their level of importance ranging from their level of importance ranging
this) and 1(keep this) for each
-1 to 1. from -1 to 1 and multiplied with an
number in the cell state Ct-1.
output of sigmoid.
LSTM Code
CNN in Computer Vision – Image
Classification
Collab
Notebooks
 Link to all collab
notebooks related
to each and every
chapter:
https://drive.google.
com/drive/folders/1y
yYYYG4KRmZMokVJ9
8PW9_x33jt8P7nq?us
p=drive_link
Final Project:
Text Summarisation

 Link:
https://colab.research.google.co
m/drive/1ZOF3XtRw31KvMQSww
mn1a4ALCChyOhpZ