Artificial Intelligence

Artificial Intelligence
1. Artificial Intelligence is a combination of 2 words :-
(i) Artificial - Things made by humans
(ii) Intelligence - The ability to think, understand and learn
2. AI makes it possible for a machine to think, understand, learn and
work like human beings.
3. When a machine possesses the ability to mimic human traits, that is,
make decisions, predict the future, learn and improve on its own, it
is said to have artificial intelligence

Intelligence
1. It is the ability to think, understand and learn
2. It is also the ability to perceive or infer information, and to retain
it as knowledge to be applied towards adaptive behaviours within an
environment or context.

Types of Intelligence
1. Mathematical Logical Reasoning - A person's ability to regulate,
measure, and understand numerical symbols, abstraction and
logic.
2. Linguistic Intelligence – Language processing skills both in terms
of understanding or implementation in writing or verbally.
3. Spatial Visual Intelligence – It is defined as the ability to
perceive the visual world and the relationship of one object to
another.
4. Kinaesthetic Intelligence – Ability that is related to how a
person uses his limbs in a skilled manner
5. Musical Intelligence – A person's ability to recognize and create
sounds, rhythms, and sound patterns
6. Existential Intelligence – An additional category of intelligence
relating to religious and spiritual awareness.
7. Naturalist Intelligence – An additional category of intelligence
relating to the ability to process information on the environment
around us
8. Intrapersonal Intelligence – Describes how high the level of
self-awareness someone has is, starting from realizing weakness,
strength, to his own feelings.
9. Interpersonal Intelligence - Ability to communicate with others
by understanding other people's feelings & influence of the
person

Machine Learning
1. It is a subset of AI which enables machines to improve at tasks
with experience (data).
2. The intention of Machine Learning is to enable machines to learn
by themselves using the provided data and make accurate
Predictions or Decisions.
Deep Learning
1. It is a branch of machine learning
2. It enables software to train itself to perform tasks with vast
amounts of data.
3. In Deep Learning, the machine is trained with huge amounts of
data which helps it in training itself around the data.
4. Such machines are intelligent enough to develop algorithms for
themselves

Data Science
1. Extracting meaningful information from the huge amount of data
through various different algorithms is called Data Science
2. It is a concept to unify statistics, data analysis, machine learning
and their related methods in order to understand and analyse
actual phenomena with data.
3. It employs techniques and theories drawn from many fields within
the context of Mathematics, Statistics, Computer Science, and
Information Science.

Computer Vision
1. Computer Vision is the domain of AI which enables machines to
see through images or visual data, process and analyse them on
the basis of algorithms and methods in order to analyse actual
phenomena with images.
2. It is the study where the computers are made to 'see' and
'identify' different objects.
 3. Identification is based on the data set provided and well labelled
images

Natural Language Processing

1. NLP is a domain of AI which is concerned with giving computers
the ability to understand text and spoken words in much the same
way human beings can
2. Its ultimate objective is to read, decipher, understand, and make
sense of the human languages in a manner that is valuable

AI Ethics
1. The concerns and principles of human beings related to good
manners or good behaviour are known as ethics
2. As human beings, we follow these moral principles for doing
certain activities and to make life comfortable
3. Similarly, certain ethics are associated with AI systems and tools
4. These ethics are called AI Ethics
5. It is of 2 types :-
(i) AI Bias
(ii) AI Access

AI Bias
1. It is a phenomenon that occurs when an algorithm produces
results that are systemically influenced due to inaccurate
assumptions in the machine learning process
2. A machine has no emotions, thus, it can’t be biased
 3. Any biasness is always transferred from the developer
4. Machine takes decisions based on data, but a developer puts
algorithms on the data to decide how it will be read

AI Access
1. AI Access is a set of ethical concerns related to the
implementation of AI systems
2. These concerns are as follows :-
(i) Unemployment
(ii) Increasing Inequalities
(iii) Negative Adoptions
(iv) Black Box Problem

Data Privacy
1. Data is our new gold
2. AI revolves around data
3. It is required by every organisation
4. Data comes from smartphones that have sensors to collect data,
social media feed which we search, like and follow
5. We, unknowingly, are providing data to websites/people
6. This data can be sold, used or even misused in many ways
7. Thus, its protection is very important
8. Accepting cookies, liking a post or even searching should be
monitored so that the data is protected
AI Project Cycle
1. It is a step by step process to solve the problems using proven
scientific methods and drawing the conclusion about it
2. It consists of 5 stages :-
(i) Problem Scoping
(ii) Data Acquisition
(iii) Data Exploration
(iv) Data Modelling
(v) Evaluation

Problem Scoping
1. It is the 1st stage in AI Project Cycle
2. It means understanding a problem, finding out different factors
which affect the problem and define the goal or aim of the
project
3. 4W Canvas :-
Who 1. ‘Who’ are facing the problem
2. All other stakeholders involved in the problem
3. Know about them
What 1. ‘What’ part helps us in understanding and
identifying the nature of the problem and under
this block, you also gather evidence to prove that
the problem you have selected actually exists
Where 1. ‘Where’ does the problem arises, situation and
the location
Why 1. ‘Why’ is the given problem worth solving
 4. Problem Statement Template :- It helps us to summarise all the
key points into one single template so that in future, whenever
there is a need to look back at the basis of the problem, we can
take a look at this and understand the key elements of it
Our [stakeholder(s)] Who
Has/have a [issue, problem, need] What
problem that
When/while [context, situation] Where
An idea [benefit of solution for them] Why
solution would

Data Acquisition
1. It is the 2nd stage of AI Project Cycle
2. It is the process of collecting accurate and reliable data to work
with
3. Data can be in the form of text, video, images and audio
4. Data Sources :-
(i) Surveys (offline)
(ii) Web Scraping
(iii) Sensors (offline)
(iv) Cameras
(v) Observations (offline)
(vi) Application Programming Interface (API)
(vii) Interview (offline)
(viii) Open-sourced Government Portals [online]
(ix) Reliable Websites (Kaggle) [online]
(x) World Organisations’ open-sourced statistical websites
[online]
5. While accessing data from any of the data sources, following
points should be kept in mind :-
 (i) Data which is available for public usage only should be taken
up.
(ii) Personal datasets should only be used with the consent of the
owner.
(iii) One should never breach someone’s privacy to collect data.
(iv) Data should only be taken form reliable sources as the data
collected from random sources can be wrong or unusable.
(v) Reliable sources of data ensure the authenticity of data
which helps in proper training of the AI model.
6. Application Programming Interface (API) :-
(i) It is a message which takes requests from the user and
then tells the system what you want and then reverts
back the response to the user.
7. Data features refer to the type of data the user wants to
collect.

Data Exploration
1. It is the 3rd stage of AI Project Cycle
2. It is the process of arranging the gathered data uniformly for a
better understanding.
3. It can be arranged in the form of a :-
(i) Table
(ii) Plotting a chart
(iii) Making a database
4. Data Exploration/Visualisation Tools :-
(i) Tableau
(ii) Microsoft Excel
(iii) Power BI
(iv) Google Charts
 5. To analyse the data, we need to visualise it in some user-friendly
format to :-
(i) Quickly get a sense of the trends, relationships and patterns
contained within the data.
(ii) Define strategy for which model to use at a later stage.
(iii) Communicate the same to others effectively.
(iv) To visualise data, we can use various types of visual
representations.

Data Modelling
1. It is the 4th stage in AI Project Cycle
2. After exploring and organising the data, we need models to
arrange the accessed data
3. After exploring the patterns, we decide which type of model fits
in best, to give us the required results
4. Types of Data Models :-
(i) Relational Data Model
(ii) Semi-Structured Data Model
(iii) Entity-Relationship Data Model
(iv) Object-Based Data Model
5. Types of AI Models :-
(i) Ruling Based
(ii) Learning Based – Machine Learning
- Deep Learning
6. Rule Based Approach/Model :-
(i) Refers to the AI modelling where the rules are defined by
the developer.
(ii) The machine follows the rules or instructions mentioned by
the developer and performs its task accordingly.
(iii) A drawback/feature for this approach is that the learning is
static
(iv) The machine once trained, does not take into consideration
any changes made in the original training dataset
(v) It follows an algorithm or code to train, test and validate
data
(vi) Training Set – The data where the model is trained on
(vii) Validation Set – When the model hasn’t been trained on and
used to tune hyperparameters
(viii) Testing Set – Same like the validation set, just used at the
final end after the model has been tailored
7. Learning Based Approach :-
(i) Refers to the AI modelling where the machine learns by itself.
(ii) Under the Learning Based approach, the AI model gets trained
on the data fed to it and then is able to design a model which is
adaptive to the change in data
(iii) The machine learns by itself by adapting to the new data which
is flowing in.
(iv) This is the machine learning approach which introduces the
dynamicity in the model.
(v) It can be divided into 3 parts :-
a) Supervised Learning – Classification, Regression, Decision
Tree
b) Unsupervised Learning – Clustering, Dimensionality Reduction
c) Reinforcement Learning
Supervised Learning
1. In a supervised learning model, the dataset which is fed to the
machine is labelled.
2. 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
3. A label is some information which can be used as a tag for data
4. Eg - students get grades according to the marks they secure in
examinations and these grades are labels which categorise the
students according to their marks
5. Classification :-
(i) Where the data is classified according to the labels
(ii) The entries are divided in 2 classes normally
(iii) The boundary condition is defined to classify

(iv)
6. Regression :-
(i) Deals with continuous data
(ii) It is linear as well as non-linear
(iii) Eg – if we know the growth rate, we can predict the salary
of someone after a certain number of years

(iv)
 7. Decision Tree :-
(i) Decision Trees in AI are used to arrive at conclusions
based on the data available from decisions made in the
past
(ii) This algorithm can be used for regression and
classification problem

Unsupervised Learning
1. An unsupervised learning model works on unlabelled dataset.
2. 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.
3. The unsupervised learning models are used to identify
relationships, patterns and trends out of the data which is fed
into it.
4. It helps the user in understanding what the data is about and
what are the major features identified by the machine in it.
5. Clustering :-
(i) Refers to the unsupervised learning algorithm which can
cluster the unknown data according to the patterns or trends
identified out of it.
(ii) 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.

(iii)
 6. Dimensionality Reduction :-
(i) We humans are able to visualise upto 3-Dimensions only but
according to a lot of theories and algorithms, there are various
entities which exist beyond 3-Dimensions
(ii) If we have a ball in our hand, it is 3-D right now
(iii) But if we click its picture, the data transforms to 2-D
(iv) Hence, to reduce the dimensions and still be able to make sense
out of the data, we use dimensionality reduction

Reinforcement Learning
1. It is defined as a machine learning method that is concerned with
how software agents should take actions in an environment
2. It is a part of a deep learning method that helps you to maximize
some portion of the cumulative reward

Evaluation
1. It is the 5th stage in AI Project Cycle
2. Evaluation is the process of understanding the reliability of any
AI model, based on outputs by feeding test dataset into the
model and comparing with actual answers.
3. There can be different Evaluation techniques, depending of the
type and purpose of the model.
4. It is not recommended to use the data we used to build the
model to evaluate it.
5. This is because our model will simply remember the whole
training set, and will therefore always predict the correct label
for any point in the training set.
6. This is known as overfitting.
7. The efficiency of the model is calculated on the basis of :-
(i) Accuracy
(ii) Precision
(iii) Recall
(iv) F1 Score
8. Accuracy :-
(i) Accuracy is defined as the percentage of correct
predictions out of all the observations

(ii)
9. Precision :-
(i) Precision is defined as the percentage of true positive
cases versus all the cases where the prediction is true

(ii)
10. Recall :-
(i) Recall is defined as the fraction of positive cases that
are correctly identified

(ii)
11. F1 Score :-
(i) F1 score can be defined as the measure of balance
between precision and recall.

(ii)
Confusion Matrix
1. The result of comparison between the prediction and reality can
be recorded in what we call the confusion matrix.
2. The confusion matrix allows us to understand the prediction
results
3. TP :-
(i) True Positive
(ii) Prediction and Reality Matches (True)
(iii) Prediction is True (Positive)
4. TN :-
(i) True Negative
(ii) Prediction and Reality Matches (True)
(iii) Prediction is False (Negative)
5. FP :-
(i) False Positive
(ii) Prediction and Reality DO NOT match (False)
(iii) Prediction is True (Positive)
6. FN :-
(i) False Negative
(ii) Prediction and Reality DO NOT match (False)
(iii) Prediction is False (Negative)

7.
Neural Networks
1. Neural Networks are a series of networks of independent neurons
just like in our brain, but in computers, neurons are not the living
tissues.
2. It is just an algorithm which give output based on the given data
3. It is essentially a system of organizing machine learning
algorithms to perform certain tasks.
4. It is a fast and efficient way to solve problems for which the
dataset is very large, such as in images.
5. Processing Layers :-
(i) Input Layer
(ii) Hidden Layer
(iii) Output Layer
6. Input Layer :-
(i) It is the 1st layer of neural networks
(ii) It is composed of artificial input neurons
(iii) It brings the initial data into the system for further
processing by subsequent layers of artificial neurons
(iv) There can be MANY input layers
7. Hidden Layer :-
(i) It is the 2nd layer of neural networks
(ii) It is a layer in between input and output layers
(iii) It is where artificial neurons take in a set of weighed
inputs and produce an output through an activation
function
(iv) It is like the brain in a human being
8. Output Layer :-
(i) It is the 3rd layer of neural networks
(ii) It is responsible for producing the final result
(iii) There is ONLY 1 output layer
 (iv) It takes in the inputs which are passed in from the layers
before it, performs the calculation via its neurons and
then the output is computed
9. Features of Neural Networks :-
(i) Neural network systems are modelled on the human brain
and nervous system
(ii) They are able to automatically extract features without
input from the programmer
(iii) Every neural network node is essentially a machine learning
algorithm
(iv) It is useful when solving problems for which the data set is
very large

Data Collection
1. Data collection is an exercise which does not require even a tiny
bit of technological knowledge.
2. But when it comes to analysing the data, it becomes a tedious
process for humans as it is all about numbers and alpha-numerical
data.
3. This is where Data Science comes into the picture.
4. It not only gives us a clearer idea around the dataset, but also
adds value to it by providing deeper and clearer analyses around
it.
5. And as AI gets incorporated in the process, predictions and
suggestions by the machine become possible on the same
6. Eg – banks, ATM machines, theatres etc
 7. Banks :-
(i) Databases of loans issued
(ii) Account holder
(iii) Locker owners
(iv) Employee registrations
(v) Bank visitors
8. ATM Machines :-
(i) Usage details per day
(ii) Cash denominations transaction details
(iii) Visitor details
9. Theatres :-
(i) Movie details
(ii) Tickets sold offline
(iii) Tickets sold online
(iv) Refreshment purchases

Types of Data
1. CSV :-
(i) CSV stands for comma separated values.
(ii) It is a simple file format used to store tabular data.
(iii) Each line of this file is a data record and reach record
consists of one or more fields which are separated by
commas.
(iv) Since the values of records are separated by a comma, hence
they are known as CSV files.
2. Spreadsheet :-
(i) A Spreadsheet is a piece of paper or a computer program
which is used for accounting and recording data using rows
and columns into which information can be entered.
 (ii) Microsoft excel is a program which helps in creating
spreadsheets.
3. SQL :-
(i) SQL is a programming language also known as Structured
Query Language.
(ii) It is a domain specific language used in programming and is
designed for managing data held in different kinds of DBMS
(Database Management System)
(iii) It is particularly useful in handling structured data.

Classification
1. Image Classification problem is the task of assigning an input
image one label from a fixed set of categories.
2. This is one of the core problems in CV that, despite its simplicity,
has a large variety of practical applications.

Classification + Localization
1. This is the task which involves both processes of identifying what
object is present in the image and at the same time identifying at
what location that object is present in that image.
2. It is used only for single objects.

Objection Detection
1. Object detection is the process of finding instances of real-world
objects such as faces, bicycles, and buildings in images or videos.
 2. Object detection algorithms typically use extracted features and
learning algorithms to recognize instances of an object category.
3. It is commonly used in applications such as image retrieval and
automated vehicle parking systems.

Instance Segmentation
1. Instance Segmentation is the process of detecting instances of
the objects, giving them a category and then giving each pixel a
label on the basis of that.
2. A segmentation algorithm takes an image as input and outputs a
collection of regions (or segments).

Pixel
1. The word “pixel” means a picture element.
2. Every photograph, in digital form, is made up of pixels.
3. They are the smallest unit of information that make up a picture
4. Usually round or square, they are typically arranged in a 2-
dimensional grid.

Resolution
1. The number of pixels in an image is sometimes called the
resolution.
2. When the term is used to describe pixel count, it can be
expressed in 2 conventions :-
 (i) Width by the Height (a monitor resolution of 1280×1024.
This means there are 1280 pixels from one side to the
other, and 1024 from top to bottom)
(ii) Single Number (pixels along the width multiplied by the
pixels along the height of the image taken by the camera)

Pixel Value
1. Each of the pixels that represents an image stored inside a
computer has a pixel value which describes how bright that pixel
is, and/or what colour it should be.
2. The most common pixel format is the byte image, where this
number is stored as an 8-bit integer giving a range of possible
values from 0 to 255.
3. Typically, zero is to be taken as no colour or black and 255 is
taken to be full colour or white.
4. In the computer systems, computer data is in the form of ones
and zeros, which we call the binary system.
5. Each bit in a computer system can have either a zero or a one.
6. Each pixel uses 1 byte of an image, which is equivalent to 8 bits of
data.
7. Since each bit can have two possible values which tells us that,
the 8 bits can have 255 possibilities of values which starts from
0 and ends at 255.
Grayscale Images
1. Grayscale images are images which have a range of shades of gray
without apparent colour.
2. The darkest possible shade is black, which is the total absence of
colour or 0 value of pixel.
3. The lightest possible shade is white, which is the total presence
of colour or 255 value of a pixel.
4. Intermediate shades of gray are represented by equal brightness
levels of the three primary colours.

RGB Images
1. All the images that we see around are coloured images.
2. These images are made up of three primary colours Red, Green
and Blue.
3. All the colours that are present can be made by combining
different intensities of red, green and blue.
4. Every RGB image is stored in the form of three different
channels called the R channel, G channel and the B channel.
5. Each plane separately has a number of pixels with each pixel value
varying from 0 to 255.
6. All the three planes when combined together form a colour image.
7. This means that in a RGB image, each pixel has a set of three
different values which together give colour to that particular
pixel.
Chatbots
1. Chatbots are software applications that use AI and NLP to
understand what a human wants and guides them to their desired
outcome with as little work for the end user as possible
2. Difference between script bot and smart bot :-
SCRIPT BOT SMART BOT
1. Script bots are easy to 1. Smart-bots are flexible and
make powerful
2. Script bots work around a 2. Smart bots work on bigger
script which is programmed databases and other resources
in them directly
3. Mostly they are free and 3. Smart bots learn with more
are easy to integrate to a data
messaging platform
4. No or little language 4. Coding is required to take
processing skills this up on board
5. Limited functionality 5. Wide functionality

Human Language vs Computer Language

1. Human Language :-
(i) Humans communicate through language which we process all
the time.
(ii) Our brain keeps on processing the sounds that it hears
around itself and tries to make sense out of them all the
time
(iii) The sound reaches the brain through a long channel.
(iv) As a person speaks, the sound travels from his mouth and
goes to the listener’s eardrum.
 (v) The sound striking the eardrum is converted into neuron
impulse, gets transported to the brain and then gets
processed.
(vi) After processing the signal, the brain gains understanding
around the meaning of it.
(vii) If it is clear, the signal gets stored.
(viii) Otherwise, the listener asks for clarity to the speaker.
(ix) This is how human languages are processed by humans.
2. Computer Language :-
(i) The computer understands the language of numbers.
(ii) Everything that is sent to the machine has to be converted
to numbers.
(iii) And while typing, if a single mistake is made, the computer
throws an error and does not process that part.
(iv) The communications made by the machines are very basic and
simple

Syntax
1. Syntax refers to the grammatical structure of a sentence.

Corpus
1. The term used for the whole textual data from all the documents
altogether is known as corpus
Text Normalisation
1. Text Normalisation helps in cleaning up the textual data in such a
way that it comes down to a level where its complexity is lower
than the actual data
2. It has the following steps :-
(i) Sentence Segmentation
(ii) Tokenisation
(iii) Removing Stopwords, Special Characters and Numbers
(iv) Converting text to a common case
(v) Stemming
(vi) Lemmatization

Sentence Segmentation
1. Under sentence segmentation, the whole corpus is divided into
sentences.
2. Each sentence is taken as a different data so now the whole
corpus gets reduced to sentences

Tokenization
1. After segmenting the sentences, each sentence is then further
divided into tokens.
2. Tokens is a term used for any word or number or special
character occurring in a sentence.
3. Under tokenisation, every word, number and special character is
considered separately and each of them is now a separate token.
Removing Stopwords, Special Characters and
Numbers
1. In this step, the tokens which are not necessary are removed
from the token list
2. Stopwords are the words which occur very frequently in the
corpus but do not add any value to it.
3. Humans use grammar to make their sentences meaningful for the
other person to understand.
4. But grammatical words do not add any essence to the information
which is to be transmitted through the statement hence they
come under stopwords
5. Few examples are - a, an, and, are, as, for, it, is, into, in, if, on, or,
such, the, there, to etc
6. Note that no tokens have to be removed if we have very little
data and since the frequency of all the words is almost the same,
no word can be said to have lesser value than the other.

Converting text to a common case

1. After the stopwords removal, we convert the whole text into a
similar case, preferably lower case.
2. This ensures that the case-sensitivity of the machine does not
consider same words as different just because of different
cases.
Stemming
1. In this step, the remaining words are reduced to their root
words.
2. In other words, stemming is the process in which the affixes of
words are removed and the words are converted to their base
form.
3. Stemming does not take into account if the stemmed word is
meaningful or not.
4. It just removes the affixes hence it is faster.
5. Few examples :-
Word Affixes Stem
Healed -ed Heal
Studies -ies Studi
Studying -ing Study

Lemmatization
1. In this step, the remaining words are reduced to their root
words.
2. In other words, lemmatization is the process in which the affixes
of words are removed and the words are converted to their base
form
3. Lemmatization makes sure that lemma is a word with meaning and
hence it takes a longer time to execute than stemming.
4. Few examples :-
Word Affixes Lemma
Healed -ed Heal
Healing -ing Heal
Studies -ies Study
Studying -ing Study
Bag of Words
1. Bag of Words is an NLP model which helps in extracting features
out of the text which can be helpful in machine learning
algorithms.
2. In bag of words, we get the occurrences of each word and
construct the vocabulary for the corpus.
3. It gives us 2 things :-
(i) A vocabulary of words for the corpus
(ii) The frequency of these words
4. Step by step approach to implement it :-
(i) Text Normalisation: Collect data and pre-process it
(ii) Create Dictionary: Make a list of all the unique words
occurring in the corpus. (Vocabulary)
(iii) Create document vectors: For each document in the
corpus, find out how many times the word from the
unique list of words has occurred.
(iv) Create document vectors for all the documents
Applications
AI DS CV NLP
1. Google 1. Fraud and Risk 1. Facial Recognition 1. Automatic
2. Siri Detection 2. Face Filters Summarization
3. Google Maps 2. Genetics and 3. Google's Search 2. Sentiment
4. FIFA Genomics by Image Analysis
5. E-Sports 3. Internet Search 4. CV in Retail 3. Text
6. Amazon 4. Targeted 5. Self-Driving Cars Classification
7. Netflix Advertising 6. Medical Imaging 4. Virtual
8. Spotify 5. Website 7. Google Translate Assistants
9. Instagram Recommendations App
10. Chatbots 6. Airline Route
11. Robots Planning
12. Humanoids
(Sofia is the
very first
humanoid
robot)