Artificial Intelligence

Artificial intelligence (AI) is a method that enables a machine to carry out

all cognitive tasks that would typically be performed by humans, such as
perceiving, learning, and reasoning.

“The Science and Engineering of making intelligent machines, especially

intelligent Computer programs
is Artificial intelligence” –JOHN MC CARTHY [Father of AI]
World Famous AI Machines
IBM Watson
An IBM supercomputer called Watson uses artificial intelligence (AI) with
sophisticated analytical software for the best results as a “question-
answering” machine.

Chatbot – Alexa, Sire, Google’s Home

A chatbot is a software program or a computer programme that mimics
human conversation through voice or text exchanges. More users are using
chatbot virtual assistants to complete simple tasks in business-to-business
(B2B) and business-to-consumer (B2C) situations.

Boston Dynamics AI Robot

Boston Dynamics focuses on developing robots with superior dexterity,
intelligence, and mobility. Boston Dynamics has introduced the second of
its industrial robots. Stretch, a commercially produced warehouse box-
moving robot, is now on display after the company introduced Spot, a four-
legged robot dog, on the market in 2020.

History of AI
Artificial Intelligence In the year 1950, Since Alan Turning established
the “Turning Test” to measure intelligence in the 1950s of the previous
century, the field of modern artificial intelligence has gained momentum.
Artificial Intelligence In the year 1955, The term “artificial intelligence”
was first used by John McCarthy, considered the father of artificial
intelligence. McCarthy has made the most to modern artificial intelligence,
along with Alan Turing, Allen Newell, Herbert A. Simon, and Marvin Minsky.
Artificial Intelligence In the year 1970, During the 1970s, the computer
era expanded. These devices were more efficient, less expensive, and
could store more data. They were incredibly capable of abstract thought,
self-recognition, and natural language processing.
Artificial Intelligence In the year 1980, The funding for research and
algorithmic tools came over these years. A deeper user experience
improved computers and boosted learning abilities.
 Artificial Intelligence In the year 2000, Many failed attempts,
unfortunately! By the year 2000, the technique had become widely
accepted. The milestones were recognised as things that needed to be
done. Despite a lack of government funding and popular support, AI might
yet succeed.
Artificial Intelligence Based Application
Gmail – automatically separating emails into “Spam” and “Not Spam”
categories. Your time is greatly reduced by spam emails being
automatically sent to the spam folder.
YouTube – YouTube will suggest videos to view based on their subject
and, to a large extent, these suggestions will match the videos you have
selected.
Flipkart or Amazon – You’re being advised to purchase products of your
choice by Flipkart or Amazon.
Difference between Conventional
programming and Machine Learning
While both traditional programming and machine learning (ML) coding are
computer programmes, their approaches and goals are different. Like your
school uniform and your dress casual, both are made of fabric but serve
different purposes.

Conventional Programming – In traditional programming, a human (the

programmer) creates the programme by hand. However, since the logic is
not programmed, rules must be manually created or manually coded.

Take a look at an example. Below are the steps to convert Celcius

scale to Fahrenheit scale
Step -1: Take input (Celcius)
Step-2: Apply the conversion formula: Fahrenheit = Celcius * 1.8 + 32
Step -3: Print the Output (Fahrenheit)
Machine Learning – machine learning, the algorithm automatically
creates the rules from the data. Data preparation, natural language user
interfaces, automatic outlier detection, recommendations, causality and
importance recognition, and many more areas can all see a rise in the
value of your embedded analytics. All of these characteristics contribute to
accelerating user insights and lowering bias in decisions.
For example, if the same Python program above is to be written using
the Machine Learning approach, the code will look like this:
Step 1: Feed lot many values in Celcius (i.e. -40, -10, 0, 8, 15, 22, 38)
Step -2: Feed corresponding Fahrenheit values (i.e. -40, 14, 32, 46, 59, 72,
100)
Step -3: Pass these 2 sets of values to Machine Learning (ML) algorithm
Step- 4: Now you ask the ML program to predict (convert) any other celcius
value to Fahrenheit, and program will tell you the answer.
How is machine learning related to AI?
AI is a field of study that aims to develop intelligent computers with human-
like abilities, such as speech recognition, vision, information assimilation,
planning, and problem-solving. In general, AI encompasses all disciplines
or technologies that seek to build intelligent machines.

Without being formally programmed, machine learning enables machines

to learn, forecast, and advance on their own. To put it simply, machine
learning is all about learning. A typical ML system begins in a “slow state”
(similar to that of a toddler) and eventually becomes “superior” by learning
from examples (like an adult).
What is Data? Define it.
Data is a representation of information that can be processed or
transmitted by humans or machines. Examples include information about
students, schools, sports teams, businesses, and animals.
A collection of information, such as statistics, words, images, photos, audio
or video clips, maps, measurements, observations, or even just a simple
description of something, is called data.
Type of Data
1. Structured Data
2. Unstructured Data
Structured Data
The most common kind of “structured data” is “quantitative data,” and most
of us deal with this kind of data on a daily basis. Structured data contains
established data types and formats that make it easy to insert into
database columns and spreadsheet fields. They are well-organized and
simple to analyse.
Examples of structured data is name, age, address etc.
Unstructured Data
The most common definition of “unstructured data” is qualitative data,
which cannot be handled or analysed using traditional relational database
(RDBMS) techniques.
Examples of unstructured data include text, video, audio, mobile activity,
social media activity, satellite imagery, surveillance imagery and the list
goes on.

Terminology and Related Concepts

Machine Learning
“Machine learning is the science of getting computers to act without being
explicitly programmed.” – Stanford University
“Machine learning algorithms can figure out how to perform important tasks
by generalizing from examples.” – University of Washington
The term “machine learning” (ML) is now used to refer to an application of
AI that gives the system the capacity to learn from experience and advance
utilising the data at its disposal.

Supervised, Unsupervised and

Reinforcement learning
Machine learning is often divided into three categories – Supervised,
Unsupervised and Reinforcement learning.
 Supervised Learning
An strategy to developing artificial intelligence ( AI) known as supervised
learning involves training a computer system on input data that has been
labelled for a certain output.

Unsupervised Machine Learning

An method known as unsupervised learning discovers patterns in untagged
data. The idea is to drive the machine to create a concise internal model of
its environment through imitation, which is a key learning strategy for
humans, and then draw creative inspiration from it.

Reinforcement Machine Learning

A machine learning training method called reinforcement learning rewards
desired behaviours and/or penalises undesirable ones. A reinforcement
learning agent can typically perceive and comprehend its surroundings, act,
and learn by making mistakes.

Deep Learning and Neural Networks

A neural network is an artificial intelligence technique that instructs
computers to analyse data in a manner modelled after the human brain. It
is a kind of artificial intelligence technique known as deep learning that
makes use of interconnected neurons or nodes in a layered structure to
mimic the human brain.

Artificial neural networks used in deep learning are modelled after brain-like
neural networks. The theory behind ANN in deep learning is that the human
brain forms the correct connections to carry out its functions, and that this
pattern can be replicated using silicon and wires in place of living neurons.

Artificial Neural Network

 Artificial neural networks (ANNs) are layers of computer programme
components called neurons (also known as nodes), coupled to other
neurons in a layered fashion. Until they can categorise the data as an
output, these networks transform the data from one neuron to another.
Another method to create a computer programme that learns from data is
the neural network.

The three distinct nodes known as input, hidden, and output make up the
neural network structure that is used the most frequently.

Input Node – The input node is the layer of the neural network where
information or initial data from the outside world is entered. After that, the
data is sent to the concealed node, where calculations can start.
Hidden Node – At this point, there is no link to the outside world. The
machine uses the data it obtained from the input node at this stage to
perform computation and processing on it. More than one concealed layer
is possible.
 Output Node – The final step is the output node, when computations are
completed and data is made available to the output layer for subsequent
transport back into the physical world.
Deep Learning
Deep learning is a subset of machine learning that is entirely based on
artificial neural networks. Since neural networks resemble the functioning of
the human brain, deep learning is also a form of brain impersonation. We
don’t have to explicitly programme everything in deep learning. It’s crucial
to understand that not everything in deep learning needs to be explicitly
programmed.

Let us now understand the difference between Machine Learning and

Deep Learning:
MACHINE LEARNING DEEP LEARNING

Works on Large amount of

Works on small amount of Dataset for accuracy.
Dataset.

Heavily dependent on High-e

Dependent on Low-end Machine.
Machine.

Divides the tasks into sub-tasks, solves them individually and

Solves problem end to end.
finally combine the results.

Takes less time to train Takes longer time to train.

Testing time may increase. Less time to test the data.

Here are a few examples of Deep Learning at Work:
Automated Driving – To detect objects like stop signs and traffic lights
robotically, automotive experts are employing deep learning. Deep
learning is also utilised to recognise pedestrians, which lowers the
likelihood of accidents.
Aerospace and defence – Another application of deep learning is in the
identification of satellite-observed objects and the location of safe and risky
areas for troops.
Medical Research – Cancer researchers utilise deep learning to
automatically identify cancer cells.
Industrial Automation – By automatically determining when individuals or
things are too close to heavy machinery, deep learning is assisting in
enhancing worker safety around such equipment.
What machine learning can and cannot do?
Here are a few examples of Machine Learning that we use every day:
Virtual Personal Assistant – like Siri, Alexa, Google Home etc.
Predictions while commuting – like Traffic Forecasts on Google Maps.
 Video Surveillance – Modern video surveillance systems use artificial
intelligence (AI) to detect crimes before they occur. They observe people’s
strange behaviour, such as prolonged periods of inactivity, stumbling, or
snoozing on benches, etc.
Social Media Services – Facebook suggests friends to you based on your
connections with them and how frequently you visit their accounts. A list of
Facebook members who you can add as friends is provided based on
ongoing learning.
Filtering of email spam and viruses – Emails are organised in
accordance with some guidelines for email spam. Receiving emails is
managed by mail filtering, which also finds and deletes emails containing
dangerous codes like viruses, Trojans, or malware.
Product recommendations – After you shop online for a product, you
frequently receive emails from related merchants. The products are either
comparable or suit your preferences, which undoubtedly improves the
buying experience.
Online Fraud Detection – Machine learning is lending its potential to make
cyberspace a secure place by tracking monetary frauds online.