Deep learning and its applications

Dr.N.Sevugapandi., MCA., M.Phil.,Ph.D.,

Assistant Professor(GL) of Computer Science(UG & PG),
Government Arts and Science College,
Kovilpatti – 628 503.Thoothukudi District. Mobile: 9600412482.
E-Mail: [email protected].
Introduction
• "Deep learning" is the new big trend in Machine Learning. It
promises general, powerful, and fast machine learning, moving us
one step closer to AI.

• Deep learning is a machine learning technique that instructs

computers to do what comes naturally to humans: learn by example.

• deep learning is on the verge of becoming the new face of AI

technology, it’s time for everyone to start taking it seriously.

• Deep learning is a subset of machine learning, which in turn is a

subset of AI.
Introduction
• Ever wondered how Google identifies your phone gallery group images
based on their location? Well, all of this is the product of deep learning.

• Deep learning is a pivotal concept behind driverless cars, allowing them

to identify a stop sign, or to distinguish a pedestrian from a lamppost.
It is the key to voice control in consumer devices like phones, tablets,
TVs, and hands-free speakers.

• deep learning teaches the computer to do what comes naturally to

humans: learning by example. In autonomous vehicles, the computer
model learns to perform tasks by identifying the text, images, and
sounds. It is the key technology that defines driverless cars, helping
them identify a stop sign or differentiate a pedestrian from a lamppost.
 HOW DOES IT WORK?
• Deep learning works on the concept of repeated teaching. It trains
the computer so that it can understand a particular pattern and
also identifies a picture or voice. After recognizing, the computer
can automatically catch that word or voice.

• This way of learning is not much different from how we humans

learn. While we were kids we also learned to say a word by
listening to things around us. This is how we learned and now deep
learning is using this formula to teach computers.
Growing trend in machine
learning
Real-world examples of deep learning models already
used in industries:

• Autonomous vehicles
• Bots based on deep learning
• Sentiment based news aggregation
• Automated translations
• Image analysis and caption generation
• Coloring illustrations
• Language identification
• Customer experience
• Text generation
• An IDC report also indicates that the AI sector is
expanding, with worldwide spending on AI systems to be
nearly USD98 million by 2023.
• So, if your bet is in reinforcing AI as your career, do not
forget to include deep learning as your major.
 Task

• CEO of Google • CEO of Albertsons

• CEO of Microsoft • CEO of Micron
• CEO of Adobe • CEO of Netapp
• CEO of Twitter • CEO of Nokia
• CEO of
Mastercard
• CEO of Palo Alto
• CEO of Pepsi • CEO of Arista
• CEO of IBM • CEO of Novartis
This makes deep learning an extremely powerful tool for modern machine learning.
Therefore, we can say that Deep Learning is:

• A collection of statistical machine learning

techniques

• Used to learn feature hierarchies

• Often based on artificial neural networks

Fig: Face Recognition using Deep Networks
As shown in the image above Deep Learning works as
follows:

• At the lowest level, network fixates on patterns of local

contrast as important
• The following layer is then able to use those patterns of
local contrast to fixate on things that resemble eyes,
noses, and mouths
• Finally, the top layer is able to apply those facial features
to face templates.
• A deep neural network is capable of composing more
and more complex features in each of its successive
layers
Deep Learning and Machine Learning are both related to
Artificial Intelligence and Deep Learning is a subset of
Machine Learning.

Machine Learning algorithms are used to build models using

structured data usually. While Deep Learning algorithms uses
large volumes of structured as well as unstructured data to
build models. Machine Learning analyses huge volumes of
past data, learns from it and trains on it to build a model.
But the working principle of
Deep Learning is based to the
structure and functions of a
human brain.
• Neural Networks form the core of Deep Learning. Just like
the brain has neurons interconnected to each other, a
neural network also has interconnected neurons called
Nodes that are connected to each other in different
layers.
• Deep learning is a subfield of ML. Deep Learning
simulates the structure and working of a human brain.
A human brain contains numerous brain cells called
neurons.

• Each neuron is performing a particular task and every

neuron interconnected with each other in a network to
exchange the information.
• Likewise in Deep Learning, different layers of the network are
created to solve the problem.

• Deep learning algorithms are used to solve complex problems like

image, text & sound analysis.

• Deep Learning can be applied to Self Driving Car, Detecting the

language, meaning of text & sound, medical research, industrial
automation etc.

• Deep learning algorithms achieve a high degree of accuracy with

a large amount of data.
There are two main reasons why
Deep Learning is trending

1) Deep learning mimics the way organic brain sort the

information, which provides favorable results for the
complex problems.

2) With the power of high processing CPUs & GPUs, it

reduces the training time.
• Deep learning (also called deep structural
learning or hierarchical learning is a set
of algorithms in machine learning that attempt to model
high-level abstractions in data by using model architectures
composed of multiple non-linear transformations.(Source:
Wikipedia).
• Deep Learning is a subfield of machine learning concerned
with algorithms inspired by the structure and function of
the brain called artificial neural networks.
What is Deep Learning?
• Think of it as mimicking the human brain. Deep learning
involves training artificial neural networks with multiple
layers to learn complex patterns from data without
explicit feature engineering.

• Think of layers as filters. Each layer extracts higher-level

features from the previous layer's output, eventually
leading to predictions or classifications.

• Examples: Image recognition, speech recognition,

natural language processing, self-driving cars.
Top features of Deep Learning:
• Deep learning's prediction and self-defense in
cybersecurity.

• Enhanced productivity and efficiency by revamping

response times.

• Deep learning helps businesses with Full-fledged and

updated market analysis.

• Optimization of the logistic systems and processes of the

company.
Why is Deep Learning Trending?
• Handles complex data: Deep learning excels with
images, audio, and text, where feature engineering is
challenging or impractical.
• Automatic feature learning: No need for hand-crafted
features - deep networks learn them automatically
from data.
• High accuracy: Deep learning often achieves state-of-
the-art performance in various tasks.
• Adaptability: Deep networks can be adapted to a
wide range of problems with relative ease.
 Here are a few key things you
need to know:
• You need to have a good understanding of statistics.
• You need to understand how digital data is structured.
• You need to understand how machine learning works.
• You need to understand how neural networks work.
• Finally, you need to understand how to create and use
machine learning models.

• Machine learning is used in a variety of fields, including

finance, marketing, bioinformatics, and natural
language processing.
What are the different ways one
can get started with machine
learning?
• One option is to enroll in a machine learning
course offered by a local college or university.
• Another option is to take a self-paced course offered
by a software company or online course provider.
• And finally, one can use online resources like articles,
e-books, and video lectures to learn about machine
learning.
What is the difference between
deep learning, machine learning and
AI?
• Scenario :
Let’s say you are in 2199 A.D. and your mom told you to go
to market and fetch one kilogram of Alphonso mango.
Now you are too exhausted after playing a FIFA99 match
and don’t want to go anywhere. So you told your robot to
do the work for you and deliver the mangoes to you. So
you have to train your robot.
Machine Learning :
• Machine Learning is a set of learning algorithm that is
being used in today's world as state-of-the-art ways to
build an intelligent machine. This set of algorithms are
frequently used because they require less data and are
easily comprehensible than Deep Learning.
Machine Learning
• Now coming to the scenario, your robot has reached
fruit shop which has apple, mango and orange. Now
differentiating between mango from orange, apple etc.,
is machine learning. For this you can train your robot
with size(weight) and color(0 for orange, 1 for red and 2
for yellow), and your robot can fetch mangoes for you.
Deep Learning:
• Deep Learning is a sub-type of learning, but here we
take the learning to a step deeper. So you’ll need more
data and data of more attributes than machine
learning.
Deep Learning
• In this scenario, detecting Alphonso mango from other
mango varieties(which looks almost same in size and
color) is Deep Learning. Your mom said to bring Alphonso
and you can’t bring any mango to complete your job.
AI :
• AI in it’s true sense means to mimic human intelligence.
And learning is one of the steps of intelligence. What to
do with the insights gained from learning is AI.
AI
• Now after your robot picked one kilogram of Alphonso
it should be able to realize that he has completed the
job and he need to return to home with out damaging
the mangoes. After reaching home he should deliver it
to you, not to your mom so that you can get appraisal
and a good mango shake from your mom.
What is the difference between deep learning, machine
learning and AI?

காட்சி:

நீங்கள் கி.பி 2199 இல் இருக்கிறீர்கள் என்று

வைத்துக்கொள்வோம், உங்கள் அம்மா உங்களை சந்தைக்குச்
சென்று ஒரு கிலோ அல்போன்சா மாம்பழத்தை
எடுத்து வரச் சொன்னார்கள். இப்போது நீங்கள் FIFA99
போட்டியில் விளையாடிய பிறகு மிகவும் சோர்வாக
உள்ளீர்கள், மேலும் எங்கும் செல்ல விரும்பவில்லை.
எனவே உங்களுக்கான வேலையைச் செய்து மாம்பழங்களை
உங்களுக்கு வழங்குமாறு உங்கள் ரோபோவிடம் சொன்னீர்கள்.
எனவே நீங்கள் உங்கள் ரோபோவுக்கு பயிற்சி அளிக்க
வேண்டும்.
Machine Learning :

• இயந்திர கற்றல் என்பது கற்றல் வழிமுறையின்

தொகுப்பாகும், இது இன்றைய உலகில்
அறிவார்ந்த இயந்திரத்தை உருவாக்குவதற்கான
அதிநவீன வழிகளாகப் பயன்படுத்தப்படுகிறது.
இந்த அல்காரிதம்களின் தொகுப்பு அடிக்கடி
பயன்படுத்தப்படுகிறது, ஏனெனில் அவற்றுக்கு
குறைவான தரவு தேவைப்படுகிறது மற்றும்
ஆழமான கற்றலை விட எளிதில்
புரிந்துகொள்ளக்கூடியது.
Machine Learning
• இப்போது காட்சிக்கு வருகிறேன், உங்கள்
ரோபோ ஆப்பிள், மாம்பழம் மற்றும் ஆரஞ்சு
பழங்கள் உள்ள பழக் கடையை அடைந்துள்ளது.
இப்போது ஆரஞ்சு, ஆப்பிள்
போன்றவற்றிலிருந்து மாம்பழத்தை
வேறுபடுத்துவது இயந்திரக் கற்றல். இதற்காக
நீங்கள் உங்கள் ரோபோவை அளவு(எடை)
மற்றும் வண்ணத்துடன் (ஆரஞ்சுக்கு 0, சிவப்புக்கு
1 மற்றும் மஞ்சள் நிறத்திற்கு 2) பயிற்சி
செய்யலாம், மேலும் உங்கள் ரோபோ
உங்களுக்காக மாம்பழங்களை எடுத்து வரலாம்.
Deep Learning:
• ஆழ்ந்த கற்றல் என்பது ஒரு துணை வகை கற்றல்,
ஆனால் இங்கே நாம் கற்றலை ஒரு படி ஆழமாக
கொண்டு செல்கிறோம். எனவே உங்களுக்கு
இயந்திர கற்றலை விட கூடுதல் தரவு மற்றும்
கூடுதல் பண்புகளின் தரவு தேவைப்படும்.
Deep Learning
• இந்தச் சூழ்நிலையில், அல்போன்சா மாம்பழத்தை
மற்ற மாம்பழ வகைகளிலிருந்து (அளவிலும்
நிறத்திலும் கிட்டத்தட்ட ஒரே மாதிரியாகத் தெரிகிறது)
கண்டறிவது ஆழ்ந்த கற்றல். உங்கள் அம்மா
அல்போன்சாவை அழைத்து வரச் சொன்னார், உங்கள்
வேலையை முடிக்க மாம்பழம் எதுவும் கொண்டு வர
முடியாது.
AI :
• AI என்பது உண்மையான அர்த்தத்தில் மனித
புத்திசாலித்தனத்தை பிரதிபலிப்பதாகும். மேலும்
கற்றல் என்பது அறிவாற்றலின் படிகளில்
ஒன்றாகும். கற்றல் மூலம் பெறப்பட்ட
நுண்ணறிவை என்ன செய்வது AI.
AI
• இப்போது உங்கள் ரோபோ ஒரு கிலோகிராம்
அல்போன்சாவை எடுத்த பிறகு, அவர்
வேலையை முடித்துவிட்டார் என்பதை உணர
முடியும், மேலும் அவர் மாம்பழங்களை
சேதப்படுத்தாமல் வீட்டிற்கு திரும்ப
வேண்டும். வீட்டை அடைந்த பிறகு அவர்
அதை உங்களுக்கு வழங்க வேண்டும், உங்கள்
அம்மாவிடம் அல்ல, அதனால் நீங்கள் உங்கள்
அம்மாவிடம் இருந்து மதிப்பீட்டையும் நல்ல
மாம்பழ குலுக்கல்லையும் பெறலாம்.
Tools used for Deep Learning
• Python
– Keras
– Tensorflow
– Theano
– Caffe
Algorithms used for Deep Learning
• Convolutional Neural Networks (CNNs)
• Long Short Term Memory Networks (LSTMs)
• Recurrent Neural Networks (RNNs)
• Generative Adversarial Networks (GANs)
• Radial Basis Function Networks (RBFNs)
• Multilayer Perceptrons (MLPs)
• Self Organizing Maps (SOMs)
• Deep Belief Networks (DBNs)
• Restricted Boltzmann Machines( RBMs)
• Autoencoders
APPLICATION OF DEEP LEARNING

 Self Driving Cars

 Virtual Assistants
 Virtual Recognition
 Natural Language Processing
 Entertainment (Netflix, sports highlights, VEVO, etc)
 Fraud Detection
SELF DRIVING CARS
• Data from cameras, geo-mapping, sensors are assisting
in creating brief and sophisticated models to guide
through traffic, identify ways, and real-time components
such as traffic volume and road stoppages.
• Also, the significant worry for self-driving car developers
is dealing with uncommon situations. A customary
pattern of testing and execution regular to deep learning
algorithms is guaranteeing safe driving with increasingly
more openness to a great many situations.
• So, we can assume that in the future deep learning will
definitely give many more such intelligent technologies.
Virtual Assistants
• Virtual Assistants are one of the very popular applications
of deep learning. We all use virtual assistants like Alexa, Siri,
Google Assistant in our day to day life.

• Another power virtual assistants are invested in is

interpreting your speech to message, make notes for you,
and book arrangements.

• With deep learning applications like text generation and

record synopses, virtual assistants can help you in making or
sending proper email duplicates also.
Virtual Recognition
• We always face a problem when we have plenty of old pictures
but we want some selected images. So, here we faced problems
and wasted time and energy in selecting pictures.

• So, here comes deep learning that helps in arranging pictures

dependent on areas identified in photos, faces, a mix of
individuals, or as per occasions, dates, and so forth.

• Large scale image visual recognition through the deep neural

networks are improving development in this section of
advanced media management by using
convolutional neural networks, Tensorflow, and Python broadly.
Natural Language Processing
• NLP is described as the natural manipulation of normal
languages, like speech and text, by software and it is the
one that aids in perfect communication between
human language and computer language.

• The area of natural language processing is one of the

most crucial and practical applications of deep learning.
Through Deep Learning, NLP is attempting to
accomplish something very similar by preparing
machines to get linguistic nuances and frame suitable
reactions.
Entertainment (Netflix, sports highlights, VEVO, etc)

• Deep Learning AI is changing the filmmaking cycle

as cameras figure out how to examine human non-
verbal communication to soak up virtual
characters.

• Content altering and auto-content creation are

now a reality because of deep learning and its
commitment to face and pattern recognition
Other Applications
• Healthcare
• Tele Medicine
• Medical Imaging
• EHR (Electronic Health Record) – COVID 19 DATA
• Drug design
• Physiological monitoring
• Tumour detection
• Entertainment (OTT) – Browsing history
• Election Predictions
• Deep dreaming
• Detecting Developmental Delay in Children
ADVANTAGE:
• The deep learning architecture is flexible to be fitted to
new troubles in the future.
• It has the proficiency to develop new features from the
limited convenient training data sets.

DISADVANTAGE:
• Due to complicated data models, it is very expensive to
train.
• It is tough to adopt by less skilled people as there is no
basic theory to teach you in choosing the correct deep
learning tools.
DISADVANTAGE
• High computational cost
• Overfitting
• Lack of interpretability
• Dependence on data quality
• Data privacy and security
concerns
• Lack of domain expertise
• Unforeseen consequences
• Limited to the data its trained on
• Black box models
What's next after deep learning?
There are a lot of things that are next for deep learning. Instead
of thinking of moving forward in one direction, think of
expanding outward in many directions:

• Better reinforcement learning / integration of deep learning

and reinforcement learning. Reinforcement learning
algorithms that can reliably learn how to control robots, etc.

• Better generative models. Algorithms that can reliably learn

how to generate images, speech, text that humans can’t tell
apart from the real thing.
• Learning to learn and ubiquitous deep learning: algorithms that
redesign their own architecture, tune their own hyper parameters, etc.
Right now it still takes a human expert to run the learning-to-learn
algorithm, but in the future it will be easier to deploy, and all kinds of
business that don’t specialize in AI will be able to leverage deep
learning.

• Machine learning for security, security for machine learning: more

cyberattacks will leverage machine learning to make more
autonomous malware, more efficient fuzzing for vulnerabilities, etc.
More cyberdefenses will leverage machine learning to respond faster
than a human could, detect more subtle intrusions, etc. ML algorithms
from opposing camps will fool each other to carry out both attacks and
defensive actions
• Dynamic routing of activity will lead to much larger models that may use
even less computation to process a single example than current models
use today. But overall, massive amounts of computation will continue to
be key for AI; whenever we make one model use less computation we’ll
just want to run thousands of models in parallel to learn-to-learn them.

• Research will focus on making extremely robust models that almost

never make a mistake, for use in safety-critical applications.

• Deep learning will continue to spread out into general culture and we’ll
see artists and meme creators using it to do things that we never would
have anticipated. I think Alexei Efros’s lab and projects like CycleGAN are
the start of this.
Here are some useful expressions and vocabulary that
can be helpful for interviews:

1. Self-introduction:
- Hello, my name is [name].
- I am pleased to be here today.
- Thank you for this opportunity.

2. Describing experience and skills:

- I have extensive experience in [field/industry].
- I possess strong communication skills.
- I am proficient in [specific software/tools].
- I have a proven track record in [relevant area].
3. Discussing strengths and weaknesses:
- One of my strengths is my ability to [specific skill or quality].
- I am particularly skilled at [specific task or ability].
- I am working on improving my [specific weakness] by [action plan].
4. Answering behavioral questions:
- In a previous role, I faced a challenging situation where I [briefly describe
the situation] and I resolved it by [explain the actions taken].
- When working in a team, I typically [describe a positive teamwork
experience].
- I pride myself on my ability to [describe a time when you demonstrated
leadership or problem-solving skills].
5. Expressing enthusiasm and interest:
- I am passionate about [specific aspect of the industry].
- This role aligns with my career goals because [explain how the role fits into
your long-term plans].
- I am excited about the opportunity to [specific aspect of the job or
company].
6. Asking questions:
- Could you tell me more about the company culture?
- What are the main responsibilities of this role?
- How does the company support professional
development?
7. Wrap-up and thank-you:
- Thank you for having me today.
- I look forward to potentially being a part of the team.
- Please feel free to reach out if there are any further
questions.
 THANK YOU all
Special thanks to SKC
and ANJAC