Artificial Intelligence (AI): Deep Learning

This is an exclusive document to the FlevyPro community - http://flevy.com/pro
Framework Primer
Artificial Intelligence (AI): Deep
Learning
Presentation created by
CONVOLUTIONAL NEURAL NETWORK RECURRENT NEUTRAL NETWORK
Input Feature extraction and mapping Output
Image classification: Human
Hidden Layer
Output layerInput layer
Context nodes

2This document is an exclusive document available to FlevyPro members - http://flevy.com/pro
Contents
Overview
Deep Learning
Convolutional Neural Network (CNN)
Recurrent Neural Network (RNN)
Templates
The content on this page has been partially hidden.
FlevyPro members can download the full document here:
https://flevy.com/browse/flevypro/artificial-intelligence-ai-deep-learning-3250

This presentation discusses a form of AI known as Deep Learning—with a
focus on Convolutional Neural Networks and Recurrent Neural Networks
Presentation Overview
The largest tech companies and governments are all heavily investing in AI research.
As organizations invest more and more into advanced automation, we are seeing significant advancements in
Artificial Intelligence (AI) in the business world. The rise of the machines is becoming an impending reality. The
AI revolution is here. Most businesses are aware of this and see the tremendous potential of AI.
The most common type of AI is currently still Machine Learning (ML), which leverages statistical techniques to
give computer systems the ability to learn with data, without being explicitly programmed.
This presentation specifically discusses a specific type of ML called Deep Learning. Deep Learning uses
computers to create networks which simulate the way a human brain perceives, organizes, and makes decisions
from data input.
This presentation further explores the most widely used models of Deep Learning in the business world:
Examples of Deep Learning being used include:
This deck also includes slide templates for you to use in your own business presentations.
Convolutional Neural Network (CNN) Recurrent Neural Network (RNN)1 2
1
2
3
Processing handwritten material
Diagnosing health diseases from medical scans
Using radar imagery to help guide self-driving cars
4
5
6
Generating captions to images
Assessing the likelihood that a credit card
transaction is fraudulent
etc.

Contents
Overview
Deep Learning
Templates

The last couple of years have seen a dramatic increase in the popularity of
Deep Learning, an approach to AI inspired by a human brain’s activity
Deep Learning – Overview
To this day, Deep Learning models have introduced advanced solutions in the areas of
computer vision, pattern recognition, natural language processing and other.
Deep Learning is a subset of Machine Learning (ML), which, in turn, is a form of Artificial Intelligence (AI). It uses
computers to create networks which simulate the way a human brain perceives, organizes, and makes decisions from
data input.
The easiest way to think of
the relationship among Deep
Learning, Machine Learning
(ML), and AI, is to visualize
them as concentric circles.
AI, the idea that came first,
is the first and largest circle.
ML, which blossomed later,
is the next circle.
Finally, Deep Learning--
which is driving today’s AI
explosion—fits inside both of
the other circles.
Artificial
Intelligence
Early artificial
intelligence stirs
excitement
Machine
Learning
Machine
learning begins
to flourish
Deep
Learning
Deep learning
breakthroughs
drive Al boom
1950’s 60’s 70’s 80’s 90’s 2000’s 10’s
Due to its accurateness
and precision in relation to
traditional ML methods,
Deep Learning is
becoming more and more
employed in a wide area
of business areas, ranging
from Marketing, Sales,
Customer Relations, etc.
Source: An Executive’s Guide to AI, McKinsey, 2017

Being a representation of cutting-edge AI technology, Deep Learning
further develops on ML to create a superior processing method
Artificial Intelligence (AI) and Deep Learning
Deep Learning can often outperform traditional Machine Learning methods.
First, let’s clarify some key concepts: Artificial Intelligence, Machine Learning, and Deep Learning.
The ability of a machine to perform cognitive
functions we associate with human minds, such
as perceiving, learning and problem solving.
Artificial
Intelligence
Machine-learning algorithms detect patterns
and learn how to make predictions and
recommendations by processing data and
experiences, rather than by receiving explicit
programming instruction. Furthermore, the
algorithms also adapt in response to new data
and experiences to improve efficacy over time.
Machine
Learning
A type of machine learning that can process a
wider range of data resources, requires less
data preprocessing by humans, and can often
produce more accurate results than traditional
machine-learning approaches.
Deep
Learning
Deep Learning is based on interconnected layers
of software-based calculators known as “neurons”
which form a neural network.
What is Deep
Learning
based on?
The network can ingest vast amounts of input data
and process them through multiple layers that
learn increasingly complex features of the data at
each layer – and thus, the network can make a
determination about the data, learn if its
determination is correct, and use what it has
learned to make determinations about new data.
How does
Deep Learning
work?
% reduction in error rate achieved by deep learning vs. traditional
methods
25
27
41Image classification
Voice recognition
Facial recognition

In the business world, Deep Learning uses a set of different models based
on different principles to process the data and create intended results
Major Types of Deep Learning
While most businesses are at early stages of adopting Deep Learning, a growing number of
companies are using it to optimize back office or consumer-facing systems and processes.
Deep Learning can be further divided in a set of different models, which use different structures and mechanisms in
order to generate output based on input without human processing. Each of these models can be used for different
tasks and can have different business uses.
Two of the most widely used models of Deep Learning in the business world are:
CONVOLUTIONAL NEURAL NETWORK RECURRENT NEUTRAL NETWORK
Hidden Layer
Context nodes
A multilayered neutral network with a
special architecture designed to extract
increasingly complex features of the data
at each layer to determine the output
A multilayered neutral network that can
store information in context nodes,
allowing it to learn data sequences and
output a number or another sequence

Contents
Overview
Deep Learning
Templates

CNN is the most popular method for object recognition—it is a specialized
kind of neural network for processing data with grid-like topology
Convolutional Neural Network (CNN) – Overview
Because of the success of CNNs in image classification, in many areas where networks are
applied to solve complex problems, often the input data first is translated to image data.
The Convolutional Neural Network (CNN) is a multilayered neural network with a special architecture designed to extract
increasingly complex features of the data at each layer to determine the output.
CNNs are used when one has an
unstructured data set (e.g. images) and
needs to infer specific information from it.
HOW IT WORKS
CNN receives an image (for example, of
letter “A”), processing it as a collection of
pixels. In the hidden, inner layers of the
model, it identifies unique features (for
example, the individual lines making up
letter “A”). The CNN can thus classify a
different image as letter “A” if it finds in it
the unique features previously identified.
USE CASES

Due to its capacity to graphically recognize or differentiate certain
elements, CNN has been put to various uses in the business world
CNN – Example Use Cases
Almost any goods producer could use CNN for quality assurance examining products
coming out of the factory or the assembly line.
CNNs are state-of-the-art technologies for image recognition. So, the question to ask, industry by industry, is what are
they looking at, or what might they look at, that affects their business.
Below there are several examples of industry-applications of CNN.
Diagnosing health diseases from medical scans
Detecting a company logo in social media to better understand joint marketing opportunities
(e.g. pairing of brands in one product)
Understanding customer brand perception and usage through images
Detecting defective products on a production line through images
Processing print or handwritten material
Using radar imagery to help guide self-driving cars

Contents
Overview
Deep Learning
Templates

Similar in many ways to CNNs, Recurrent Neural Networks are better
trained to work with text and language rather than with images
Recurrent Neural Network (RNN) – Overview
Recurrent Neural Networks are one of the most common neural networks used in natural
language processing because of its promising results.
The Recurrent Neural Network (RNN) is a multilayered neural network that can store information in context nodes,
allowing it to learn data sequences and output a number or another sequence.
RNN is used when one is working with time-series
data or sequences (e.g. audio recording or text).
HOW IT WORKS
USE CASES
A RNN neuron receives a command that indicates
the start of a sentence. Then, the neuron receives the
word “Are” and then outputs a vector of numbers that
feeds back into the neuron to help it “remember” that
it received “Are”. The same process occurs when it
receives “you” and “free”, with the state of the neuron
updating upon receiving each word. After receiving
“free”, the neuron assigns a probability to every word
in the English vocabulary that could complete the
sentence. If trained well, the RNN will assign the
word “tomorrow” one of the highest probabilities and
will choose it to complete the sentence.
Hidden Layer
Context nodes

The RNN is now having different business applications in a variety of
industries
RNN – Example Use Cases
Due to continuous innovations, RNN can now also be appropriated to train on
non-sequential data in a non-sequential manner.
Due to its capacity to understand and learn data sequences, making decisions and generating output based on the input
data, RNN has also found a multitude of applications in business, being continuously applied in different industries
where sequential data is used.
Below there are several examples of industry-applications of RNN.
Generating analyst reports for securities traders
Providing language translations
Tracking visual changes to an area after a disaster to assess potential damage claims (in conjunction
with CNNs)
Assessing the likelihood that a credit card transaction is fraudulent
Generating captions to images
Powering chatbots that can address more nuanced customer needs and inquiries

Contents
Overview
Deep Learning
Templates

Insert headline
Artificial Intelligence vs. Machine Learning vs. Deep Learning – TEMPLATE
Insert bumper.
Artificial
Intelligence
Early artificial
intelligence stirs
excitement
Machine
Learning
Machine learning
begins to flourish
Deep
Learning
Deep learning
breakthroughs
drive Al boom
1950’s 60’s 70’s 80’s 90’s 2000’s 10’s
• Insert filler text, filler text, filler text, filler text,
filler text, filler text, filler text
filler text
• Insert filler text, filler text, filler text, filler text
filler text
filler text

Insert headline
Convolutional Neural Network (CNN) – TEMPLATE
Insert bumper.
• Insert filler text, filler text,
filler text, filler text, filler
text, filler text, filler text
text
text
filler text, filler text

Insert headline
Recurrent Neural Network (RNN) – TEMPLATE
Insert bumper
Hidden Layer
Context nodes
text, filler text, filler text
text
text
filler text, filler text

Download 100s of similar frameworks from the
FlevyPro Library:
https://flevy.com/pro/library/frameworks

Need more frameworks? Download our Complete Business Frameworks
Reference Guide, a 350+ slide compilation of 50+ frameworks, on Flevy
The Complete Business Frameworks Reference
Guide is a best selling document on Flevy. It is 350+
slides--covering 50+ common management consulting
frameworks and methodologies. A summary is
provided for each business framework.
The frameworks in this deck span across Corporate
Strategy, Sales, Marketing, Operations, Organization,
Change Management, and Finance.
This reference guide is great for those who need a
refresher on common frameworks, as well as be
introduced and learn new useful frameworks.
You can find this document here;
http://flevy.com/browse/business-document/complete-
consulting-frameworks-toolkit-644

Flevy (www.flevy.com) is the marketplace
for premium documents. These
documents can range from Business
Frameworks to Financial Models to
PowerPoint Templates.
Flevy was founded under the principle that
companies waste a lot of time and money
recreating the same foundational business
documents. Our vision is for Flevy to
become a comprehensive knowledge base
of business documents. All organizations,
from startups to large enterprises, can use
Flevy— whether it's to jumpstart projects, to
find reference or comparison materials, or
just to learn.
Contact Us
Please contact us with any questions you may have
about our company.
• General Inquiries
support@flevy.com
• Media/PR
press@flevy.com
• Billing
billing@flevy.com

1
Flevy (www.flevy.com) is the marketplace
for premium documents. These
documents can range from Business
Frameworks to Financial Models to
PowerPoint Templates.
Flevy was founded under the principle that
companies waste a lot of time and money
recreating the same foundational business
documents. Our vision is for Flevy to
become a comprehensive knowledge base
of business documents. All organizations,
from startups to large enterprises, can use
Flevy— whether it's to jumpstart projects, to
find reference or comparison materials, or
just to learn.
Contact Us
Please contact us with any questions you may have
about our company.
• General Inquiries
support@flevy.com
• Media/PR
press@flevy.com
• Billing
billing@flevy.com

Artificial Intelligence (AI): Deep Learning

More Related Content

Similar to Artificial Intelligence (AI): Deep Learning (20)

More from Flevy.com Best Practices (20)

Recently uploaded (20)

Artificial Intelligence (AI): Deep Learning