SlideShare ist ein Scribd-Unternehmen logo
1 von 54
Downloaden Sie, um offline zu lesen
Simplifying AI:
What to use when?
We read about
the use ML
models and ML
Ops, Deep
Learning and
even sentient AI
everyday...
We get the
shivers when we
see machines
behaving like
impulsive people
on Netflix
movies...
How much do we know
about them? Do we know
what they mean? And what
they mean for our
business?
Let's Uncomplicate by
keeping Netflix aside...
2020
Let's Uncomplicate by
keeping Netflix aside...
2020
Let's define the coolest
term first
WHAT
IS AI?
When a machine starts to perform cognitive
functions that we associate with human
minds such as perceiving, reasoning,
learning and problem solving - we call it
'Artificial Intelligence'.
The term was coined by Prof
John McCarthy which he
explained as 'the science
and engineering of making
intelligent machines'.
Computer Vision, Robotics,
Natural Language
Processing, Machine
Learning are all
technologies that enable AI.
At the heart of AI is
machine learning
2020
What does Machine Learning do?
Most recent advances in AI has
been done through Machine
Learning. ML Algorithms detect
patterns in large sets and learn to
make predictions and
recommendations on their own.
The coolest part is - the
ML algorithms can adapt
and improve themselves
over time as more data
becomes accessible.
2020
Types of ML Outputs Descriptive
Describe what happened. This is heavily used across
industries. E.g - LMS analytics, learner behavior.
01.
Prescriptive
Describe what will happen. This is inherently
probabilistic. e.g: Financial Monitoring.
02.
Describe What to do. e.g: Making investment
predictions or predicting learner drop-out rates.
Predictive
03.
There are three types of analytics.
Out of them Prescriptive and
Predictive are key domains of ML
outputs.
MACHINE
LEARNING
The three types of Machine Learning: Type 1
Supervised
Learning
We give the input data
We give the output variables
We label all the data
We train the algorithm to make
connections between the input
and output variables.
Then we have coffee as the
algorithm works on its own
The three types of Machine Learning: Type 2
Unsupervised
Learning
We give the input data
The algorithm figures out the
output variable on its own.
E.g: An algorithm figures out
types of probable customers
for a product from social media
chatter.
We have coffee after giving
only the input data
The three types of Machine Learning: Type 3
Reinforcement
Learning
An algorithm figures out how to
do a particular task by getting
rewarded for doing it right and
punished for doing it wrong.
E.g: If a self-driving car hits a
pedestrian you take away its
fuel. That's punishment.
We stand behind the algorithm
with a stick here.
And there's
Transfer
Learning too...
When a ML algorithm has
worked something out in a
particular domain, it is easier to
apply the same knowledge and
learn something faster in a
adjacent areas.
How do we use these algos in
actual business cases
2020
Let's start with the
Supervised Learning
Algorithms
2020
Linear Regression
THE A-Z OF MAKING IT STICK 2020
Linear Regression is a standard
method for modeling the past
relationship between independent
variables and dependent variables
to help predict future values of
the output.
This can help you understand
product-sales drivers e.g - prices
of competition, distribution, etc.
This can help you optimize price
points and estimate product-price
elasticities.
What is it? Where to use?
Logistic Regression
THE A-Z OF MAKING IT STICK 2020
An extension of Linear Regression
-Logistic regression is used for
classification tasks where the
output is always in a binary (a Yes
or No answer).
Logistic regression is used in
areas where the output variable
needs to be a yes/no answer, for
example: will someone pay off a
loan, predict if a skin lesion is
malignant or benign.
What is it? Where to use?
Linear/Quadratic Discriminant Analysis
THE A-Z OF MAKING IT STICK 2020
A Linear Discriminant Analysis
upgrades a logistic regression to
deal with non-linear problems so
that if you make changes in the
input variables they do not have
proportional changes to the
output variables.
If you are a SaaS firm, this can
help you predict the possibility of
Client churn, or even the potential
of a sales person closing a lead.
What is it? Where to use?
Decision Trees
THE A-Z OF MAKING IT STICK 2020
A highly interpretable
classification or regression model
that splits data-features into
branches at decision nodes until a
final decision output is made.
If you are building a product,
Decision Trees can help you
identify which feature is most
likely to drive sales/adoption of
the product.
What is it? Where to use?
Naive Bayes
THE A-Z OF MAKING IT STICK 2020
Classification technique that
applies Bayes theorem which
allows the probability of an event
to be calculated based on the
knowledge of factors that might
affect it.
You can use Naive Bayes to
analyze sentiments - such as
understanding your product
perception in the market or even
create classifiers to identify to
filter spam emails (Naive Bayes
can help you classify emails which
say 'I am a sad billionnaire in
Nigeria' and automatically put
them in spam.
What is it? Where to use?
Naive Bayes
THE A-Z OF MAKING IT STICK 2020
Classification technique that
applies Bayes theorem which
allows the probability of an event
to be calculated based on the
knowledge of factors that might
affect it.
You can use Naive Bayes to
analyze sentiments - such as
understanding your product
perception in the market or even
create classifiers to identify to
filter spam emails (Naive Bayes
can help you classify emails which
say 'I am a sad billionnaire in
Nigeria' and automatically put
them in spam.
What is it? Where to use?
Support Vector Machine
THE A-Z OF MAKING IT STICK 2020
Support Vector Machine is a
technique that's typically used for
classification but can be
transformed to perform
regression. It draws an optimal
division between classes
You can predict how many
patients a hospital will need to
serve in a given time period.
Support Vector Machine
algorithms are also applied in text
classification, image recognition,
face recognition, etc.
What is it? Where to use?
Random Forest
THE A-Z OF MAKING IT STICK 2020
Random Forest improves the
accuracy of a single Decision Tree
by generating multiple decision
trees and then identifying the best
output by taking a majority vote.
You can predict call volumes in
call centers and take relevant
hiring and training decisions by
using the Random Forest
algorithm.
What is it? Where to use?
AdaBoost
THE A-Z OF MAKING IT STICK 2020
Think of an AdaBoost (or Adaptive
Boosting) is iike a discerning CEO
who troubles everyone for the
best result. The AdaBoost is a
classification technique that runs
a multitude of models to come up
with a decision and weighs them
based on their accuracy.
You can predict call volumes in
call centers and take relevant
hiring and training decisions by
using the Random Forest
algorithm.
What is it? Where to use?
Gradient Boosting Trees
THE A-Z OF MAKING IT STICK 2020
A classification technique where
the decision trees come in a
sequential manner by correcting
the errors of the previous decision
trees. The final output is a
combination of the results of all
the trees.
This can be used in forecasting -
such as forecasting of demand or
inventory levels in eCommerce.
What is it? Where to use?
Simple Neural Network
THE A-Z OF MAKING IT STICK 2020
A model in which artificial neurons
make up three layers ( an input
layer, a hidden layer and an output
layer) that can be used to classify
data or find relationships between
variables in regression problems.
Predict the probability of a user
paying a certain price for a
particular service
Can be used in SaaS services, OTT
services, healthcare programs,
etc.
What is it? Where to use?
Whew!
Wasn't that
quite a bit!
Here are some quick quips on
where supervised learning is going!
For supervised learning
to run effectively - it
needs tons of labeled
data
WHICH,
FRANKLY... IS A
CHALLENGE.
That's why Tesla is using
SELF SUPERVISED LEARNING.
THE A-Z OF MAKING IT STICK 2020
Supervised learning can be time consuming and may take
months to create the right dataset to train the ML model.
SELF-SUPERVISED LEARNING IS A MACHINE LEARNING
APPROACH WHERE THE MODEL TRAINS ITSELF BY
LEVERAGING ONE PART OF THE DATA TO PREDICT
THE OTHER PART AND GENERATE LABELS
ACCURATELY.
Robotic surgeries: SSL can help in robotic surgeries by
estimating the depth of the human body and providing
better medical visuals.
Autonomous cars: SSL can estimate the roughness of the
terrain.
Where can Self Supervised Learning be effective:
Let's move on to
Unsupervised Learning
K-Means Clustering
The algorithm helps put data into a
number of groups (K) each of
which contain data of similar
characteristics.
K-Means Clustering can help you
identify customer types, customer
preferences, and even customer
personas from unstructured data
such as social media
conversations.
What is it? Where to use?
K-Means Clustering
The algorithm helps put data into a
number of groups (K) each of
which contain data of similar
characteristics.
K-Means Clustering can help you
identify customer types, customer
preferences, and even customer
personas from unstructured data
such as social media
conversations.
What is it? Where to use?
Gaussian Mixture Model
A generalized version of K-Means
Clustering, the GMM allows for
more flexibility in the size and
shape of the clusters
Gaussian Mixture Model can help
in better segmentation of
customers using less-distinct
customer characteristics.
For example, you can use the
GMM to identify the chances of
employee attrition.
What is it? Where to use?
Hierarchical Clustering
It aggregates clusters in a
hierarchical tree to form a
classification system.
You can use the Hierarchical
Clustering algorithm to divide your
loyalty program customers into
micro-segments to devise a more
personalized approach towards
customer-outreach and increase
ARR.
What is it? Where to use?
Recommender Systems
A recommender system often
uses cluster behavior prediction
to identify the important data
necessary for making a
recommendation
We did get back to Netflix didn't
we, even though we kept it aside.
Netflix uses the Recommender
system to make personalized
movie recommendations to you.
Next time you get a Korean Sci Fi
recommendation after watching a
Russian Sci Fi, just remember
what was being used.
What is it? Where to use?
Now for
mistakes
and
punishment
s....
You got it! We are talking about
Reinforcement learning here.
Some Use Cases here
Reinforcement learning can be
used to make the right trade in
time. The way you train your algo
is by punishing it every-time it
makes the wrong trade.
Using a multi-modal input model
you train your self-driving car to
avoid mistakes, potholes, hitting
pedestrians on the road, going in
the wrong road, etc. Over a period
of time through a cycle of rewards
and punishments your algo
optimizes its driving behavior.
Algorthmic trading
Optimizing driving
behaviour of self-
driving cars
And now for
Deep Learning
It's the key tech behind some fo the
coolest inventions such as Self-
Driving cars, voice control behind
phones, etc.
We are
mirroring your
neurons
I
Deep Learning is a
subset of Machine
Learning that uses
algorithms inspired by
the structure and
functions of the brain's
neural network
Deep Learning and Artificial Neurons
receive input values,
weighs these values with weights and
returns an output value.
Deep Learning works with artificial neurons. "Artificial Neurons" are
mathematical tools that
The value returned by the neuron is called an 'activation' similar to the
electrical signals that a human neuron receives from its dendrites and
decides if it needs to activate it to other neurons.
How does Deep Learning work?
In Deep Learning, an Artificial Neural Network passes information
through a series of neurons in an input layer. Think of your sensory
organs (ear, nose) taking in information and passing onto your brain.
The hidden layers at the second level processes this information and
assigns weights to it. For example, if you are trying to predict an
inventory - the weightage given to past sales history will be higher than
the week when it has been sold.
Deep Learning is used in predicting and completing complex tasks such
as predicting the weather.
Types of Deep Learning Neural Networks
Recurrent Neural Networks
Convolutional Neural Networks
Multilayered Perceptrons
The three popular types of Deep Learning Neural Networks include the
following:
Convolutional Neural
Networks
Recurrent Neural Networks
A ConvNet (employed in areas like
Computer Vision) use a range of
convolutional operations to extract
the features given an image or video.
A ConvNet or CNN is a favourite when it
comes to areas like image
classification, image extraction, videos,
etc.
The applications of CNN can be seen in
a range of industries starting from
surveillence, creative, medicine, etc.
RNN uses sequential data feeding.
Thus, the inputs in RNN include
current input as well as previous
samples. RNN is good with time series
calculations and handling sequential
data.
The application of RNN include text
predictions - if you are about to type
something in the search box and you
get a series of choices based on your
previous inputs as well as common
inputs of others, then it is RNN at play.
The applications of RNN go into
understanding human language and has
implications in Natural Language
Processing and Natural Language
Understanding.
Multilayer
Perceptrons
Multilayer perceptron is a
classical neural network used for
basic operations like data
visualization, data compression,
and encryption. It basically has a
input layer, hidden layers and
output layer
Multilayer Perceptrons are used in
areas like data visualization, data
analysis, etc.
WHOA!
WASN'T THAT
A LOT? BUT
BEFORE WE
GO... HERE'S A
SHORT
TEASER...
Sunspring is a 2016 movie written by an AI
called Benjamin. Can you guess the tech that
was used in training up the AI here?

Weitere Àhnliche Inhalte

Was ist angesagt?

How AI Benefits Sales and Marketing
How AI Benefits Sales and MarketingHow AI Benefits Sales and Marketing
How AI Benefits Sales and MarketingCall Sumo
 
How does ai impact email marketing and programmatic email advertising?
How does ai impact email marketing and programmatic email advertising?How does ai impact email marketing and programmatic email advertising?
How does ai impact email marketing and programmatic email advertising?inboxAds
 
Kavita lakhani- Advanced Program in Digital Marketing - Project Submission AP...
Kavita lakhani- Advanced Program in Digital Marketing - Project Submission AP...Kavita lakhani- Advanced Program in Digital Marketing - Project Submission AP...
Kavita lakhani- Advanced Program in Digital Marketing - Project Submission AP...Kavita Lakhani
 
The Future of CRM Marketing, 8 Big Trends and How They Will Affect You
The Future of CRM Marketing, 8 Big Trends and How They Will Affect YouThe Future of CRM Marketing, 8 Big Trends and How They Will Affect You
The Future of CRM Marketing, 8 Big Trends and How They Will Affect YouArturo Gutierrez, MBA
 
Ai in marketing: How Business & Marketing can benefit from adopting AI & ML
Ai in marketing: How Business & Marketing can benefit from adopting AI & MLAi in marketing: How Business & Marketing can benefit from adopting AI & ML
Ai in marketing: How Business & Marketing can benefit from adopting AI & MLleanstartuphh
 
The growth of artificial intelligence in e commerce (1)@
The growth of artificial intelligence in e commerce (1)@The growth of artificial intelligence in e commerce (1)@
The growth of artificial intelligence in e commerce (1)@Andolasoft Inc
 
Marketing to the Power of ONE - Salesforce Partner Forum - New York City 11.1...
Marketing to the Power of ONE - Salesforce Partner Forum - New York City 11.1...Marketing to the Power of ONE - Salesforce Partner Forum - New York City 11.1...
Marketing to the Power of ONE - Salesforce Partner Forum - New York City 11.1...Joel Book
 
Netcore Falconide Triggred and Transactional Emails
Netcore Falconide Triggred and Transactional EmailsNetcore Falconide Triggred and Transactional Emails
Netcore Falconide Triggred and Transactional EmailsNetcore Solutions
 
Parke - Siebel Magazine Cover Story
Parke - Siebel Magazine Cover StoryParke - Siebel Magazine Cover Story
Parke - Siebel Magazine Cover StoryShep Parke
 
Artificial intelligence chatbot Report
Artificial intelligence chatbot ReportArtificial intelligence chatbot Report
Artificial intelligence chatbot ReportTaslima Yasmin Tarin
 
Webinar: How to Scale AI in the world of eCommerce
Webinar: How to Scale AI in the world of eCommerceWebinar: How to Scale AI in the world of eCommerce
Webinar: How to Scale AI in the world of eCommerceSakshi Singh
 
111814_Salesforce Marketing Cloud - Partner Forum NYC
111814_Salesforce Marketing Cloud - Partner Forum NYC111814_Salesforce Marketing Cloud - Partner Forum NYC
111814_Salesforce Marketing Cloud - Partner Forum NYCJoel Book
 
Transforming the Omni-channel Customer Experience
Transforming the Omni-channel Customer ExperienceTransforming the Omni-channel Customer Experience
Transforming the Omni-channel Customer ExperienceCafeX
 
Winterberry group the state of consumer data onboarding november 2016
Winterberry group   the state of consumer data onboarding november 2016Winterberry group   the state of consumer data onboarding november 2016
Winterberry group the state of consumer data onboarding november 2016Brian Crotty
 
Mapping the Omnichannel Customer Journey: How to Engage People in a Mobile Fi...
Mapping the Omnichannel Customer Journey: How to Engage People in a Mobile Fi...Mapping the Omnichannel Customer Journey: How to Engage People in a Mobile Fi...
Mapping the Omnichannel Customer Journey: How to Engage People in a Mobile Fi...FollowAnalytics
 
Combining offline and online data to drive performance - ArabNet Riyadh 2015
Combining offline and online data to drive performance - ArabNet Riyadh 2015Combining offline and online data to drive performance - ArabNet Riyadh 2015
Combining offline and online data to drive performance - ArabNet Riyadh 2015ArabNet ME
 
Grow Your Amazon Ebay Sales get eCommerce Business Solution
Grow Your Amazon Ebay Sales get eCommerce Business Solution Grow Your Amazon Ebay Sales get eCommerce Business Solution
Grow Your Amazon Ebay Sales get eCommerce Business Solution Ali Afaq
 
Business Case_E-Commerce
Business Case_E-CommerceBusiness Case_E-Commerce
Business Case_E-CommerceGautamMandal26
 
Warply Mobile Banking solutions
Warply Mobile Banking solutionsWarply Mobile Banking solutions
Warply Mobile Banking solutionsWarply
 

Was ist angesagt? (20)

How AI Benefits Sales and Marketing
How AI Benefits Sales and MarketingHow AI Benefits Sales and Marketing
How AI Benefits Sales and Marketing
 
How does ai impact email marketing and programmatic email advertising?
How does ai impact email marketing and programmatic email advertising?How does ai impact email marketing and programmatic email advertising?
How does ai impact email marketing and programmatic email advertising?
 
Marketing to the Power of ONE!
Marketing to the Power of ONE!Marketing to the Power of ONE!
Marketing to the Power of ONE!
 
Kavita lakhani- Advanced Program in Digital Marketing - Project Submission AP...
Kavita lakhani- Advanced Program in Digital Marketing - Project Submission AP...Kavita lakhani- Advanced Program in Digital Marketing - Project Submission AP...
Kavita lakhani- Advanced Program in Digital Marketing - Project Submission AP...
 
The Future of CRM Marketing, 8 Big Trends and How They Will Affect You
The Future of CRM Marketing, 8 Big Trends and How They Will Affect YouThe Future of CRM Marketing, 8 Big Trends and How They Will Affect You
The Future of CRM Marketing, 8 Big Trends and How They Will Affect You
 
Ai in marketing: How Business & Marketing can benefit from adopting AI & ML
Ai in marketing: How Business & Marketing can benefit from adopting AI & MLAi in marketing: How Business & Marketing can benefit from adopting AI & ML
Ai in marketing: How Business & Marketing can benefit from adopting AI & ML
 
The growth of artificial intelligence in e commerce (1)@
The growth of artificial intelligence in e commerce (1)@The growth of artificial intelligence in e commerce (1)@
The growth of artificial intelligence in e commerce (1)@
 
Marketing to the Power of ONE - Salesforce Partner Forum - New York City 11.1...
Marketing to the Power of ONE - Salesforce Partner Forum - New York City 11.1...Marketing to the Power of ONE - Salesforce Partner Forum - New York City 11.1...
Marketing to the Power of ONE - Salesforce Partner Forum - New York City 11.1...
 
Netcore Falconide Triggred and Transactional Emails
Netcore Falconide Triggred and Transactional EmailsNetcore Falconide Triggred and Transactional Emails
Netcore Falconide Triggred and Transactional Emails
 
Parke - Siebel Magazine Cover Story
Parke - Siebel Magazine Cover StoryParke - Siebel Magazine Cover Story
Parke - Siebel Magazine Cover Story
 
Artificial intelligence chatbot Report
Artificial intelligence chatbot ReportArtificial intelligence chatbot Report
Artificial intelligence chatbot Report
 
Webinar: How to Scale AI in the world of eCommerce
Webinar: How to Scale AI in the world of eCommerceWebinar: How to Scale AI in the world of eCommerce
Webinar: How to Scale AI in the world of eCommerce
 
111814_Salesforce Marketing Cloud - Partner Forum NYC
111814_Salesforce Marketing Cloud - Partner Forum NYC111814_Salesforce Marketing Cloud - Partner Forum NYC
111814_Salesforce Marketing Cloud - Partner Forum NYC
 
Transforming the Omni-channel Customer Experience
Transforming the Omni-channel Customer ExperienceTransforming the Omni-channel Customer Experience
Transforming the Omni-channel Customer Experience
 
Winterberry group the state of consumer data onboarding november 2016
Winterberry group   the state of consumer data onboarding november 2016Winterberry group   the state of consumer data onboarding november 2016
Winterberry group the state of consumer data onboarding november 2016
 
Mapping the Omnichannel Customer Journey: How to Engage People in a Mobile Fi...
Mapping the Omnichannel Customer Journey: How to Engage People in a Mobile Fi...Mapping the Omnichannel Customer Journey: How to Engage People in a Mobile Fi...
Mapping the Omnichannel Customer Journey: How to Engage People in a Mobile Fi...
 
Combining offline and online data to drive performance - ArabNet Riyadh 2015
Combining offline and online data to drive performance - ArabNet Riyadh 2015Combining offline and online data to drive performance - ArabNet Riyadh 2015
Combining offline and online data to drive performance - ArabNet Riyadh 2015
 
Grow Your Amazon Ebay Sales get eCommerce Business Solution
Grow Your Amazon Ebay Sales get eCommerce Business Solution Grow Your Amazon Ebay Sales get eCommerce Business Solution
Grow Your Amazon Ebay Sales get eCommerce Business Solution
 
Business Case_E-Commerce
Business Case_E-CommerceBusiness Case_E-Commerce
Business Case_E-Commerce
 
Warply Mobile Banking solutions
Warply Mobile Banking solutionsWarply Mobile Banking solutions
Warply Mobile Banking solutions
 

Ähnlich wie Simplifying ai: What to use when?

AI and ML Trends to Impact Business 2020
AI and ML Trends to Impact Business 2020AI and ML Trends to Impact Business 2020
AI and ML Trends to Impact Business 2020Takayuki Yamazaki
 
Effectiveness and Efficiency Recognise the Value of AI & ML for Organisations...
Effectiveness and Efficiency Recognise the Value of AI & ML for Organisations...Effectiveness and Efficiency Recognise the Value of AI & ML for Organisations...
Effectiveness and Efficiency Recognise the Value of AI & ML for Organisations...Flexsin
 
Responsible Generative AI Design Patterns
Responsible Generative AI Design PatternsResponsible Generative AI Design Patterns
Responsible Generative AI Design PatternsDebmalya Biswas
 
Artificial Intelligence
Artificial IntelligenceArtificial Intelligence
Artificial IntelligenceEnes Bolfidan
 
Demystifying AI
Demystifying AIDemystifying AI
Demystifying AIAnurag Sahay
 
How to choose the right AI model for your application?
How to choose the right AI model for your application?How to choose the right AI model for your application?
How to choose the right AI model for your application?Benjaminlapid1
 
Machine Learning for Lead Qualification
Machine Learning for Lead QualificationMachine Learning for Lead Qualification
Machine Learning for Lead QualificationRosanna Garcia
 
Artificial Intelligence: Classification, Applications, Opportunities, and Cha...
Artificial Intelligence: Classification, Applications, Opportunities, and Cha...Artificial Intelligence: Classification, Applications, Opportunities, and Cha...
Artificial Intelligence: Classification, Applications, Opportunities, and Cha...Abdullah al Mamun
 
The Ultimate Guide to Machine Learning (ML)
The Ultimate Guide to Machine Learning (ML)The Ultimate Guide to Machine Learning (ML)
The Ultimate Guide to Machine Learning (ML)RR IT Zone
 
Modex Talks - AI Conceptual Overview
Modex Talks - AI Conceptual OverviewModex Talks - AI Conceptual Overview
Modex Talks - AI Conceptual OverviewModex
 
AI BI and ML.pdf
AI BI and ML.pdfAI BI and ML.pdf
AI BI and ML.pdfJayarajJoshi
 
Ai Sustainability
Ai Sustainability Ai Sustainability
Ai Sustainability prateek verma
 
Ai for everyone
Ai for everyoneAi for everyone
Ai for everyoneTrang Tran
 
Machine learning applications nurturing growth of various business domains
Machine learning applications nurturing growth of various business domainsMachine learning applications nurturing growth of various business domains
Machine learning applications nurturing growth of various business domainsShrutika Oswal
 
A Practical Guide to AI and Automation
A Practical Guide to AI and AutomationA Practical Guide to AI and Automation
A Practical Guide to AI and AutomationAccelirate Inc.
 

Ähnlich wie Simplifying ai: What to use when? (20)

AI and ML Trends to Impact Business 2020
AI and ML Trends to Impact Business 2020AI and ML Trends to Impact Business 2020
AI and ML Trends to Impact Business 2020
 
Data annotation improving customer services
Data annotation improving customer servicesData annotation improving customer services
Data annotation improving customer services
 
Effectiveness and Efficiency Recognise the Value of AI & ML for Organisations...
Effectiveness and Efficiency Recognise the Value of AI & ML for Organisations...Effectiveness and Efficiency Recognise the Value of AI & ML for Organisations...
Effectiveness and Efficiency Recognise the Value of AI & ML for Organisations...
 
Responsible Generative AI Design Patterns
Responsible Generative AI Design PatternsResponsible Generative AI Design Patterns
Responsible Generative AI Design Patterns
 
Artificial Intelligence
Artificial IntelligenceArtificial Intelligence
Artificial Intelligence
 
Demystifying AI
Demystifying AIDemystifying AI
Demystifying AI
 
How to choose the right AI model for your application?
How to choose the right AI model for your application?How to choose the right AI model for your application?
How to choose the right AI model for your application?
 
Machine Learning for Lead Qualification
Machine Learning for Lead QualificationMachine Learning for Lead Qualification
Machine Learning for Lead Qualification
 
Artificial Intelligence: Classification, Applications, Opportunities, and Cha...
Artificial Intelligence: Classification, Applications, Opportunities, and Cha...Artificial Intelligence: Classification, Applications, Opportunities, and Cha...
Artificial Intelligence: Classification, Applications, Opportunities, and Cha...
 
Atharva latest
Atharva latestAtharva latest
Atharva latest
 
The Ultimate Guide to Machine Learning (ML)
The Ultimate Guide to Machine Learning (ML)The Ultimate Guide to Machine Learning (ML)
The Ultimate Guide to Machine Learning (ML)
 
Modex Talks - AI Conceptual Overview
Modex Talks - AI Conceptual OverviewModex Talks - AI Conceptual Overview
Modex Talks - AI Conceptual Overview
 
unit_5.pdf
unit_5.pdfunit_5.pdf
unit_5.pdf
 
Salesforce Einstein: Use Cases and Product Features
Salesforce Einstein: Use Cases and Product FeaturesSalesforce Einstein: Use Cases and Product Features
Salesforce Einstein: Use Cases and Product Features
 
AI BI and ML.pdf
AI BI and ML.pdfAI BI and ML.pdf
AI BI and ML.pdf
 
Demystifying AI | A Comprehensive Guide
Demystifying AI | A Comprehensive Guide	Demystifying AI | A Comprehensive Guide
Demystifying AI | A Comprehensive Guide
 
Ai Sustainability
Ai Sustainability Ai Sustainability
Ai Sustainability
 
Ai for everyone
Ai for everyoneAi for everyone
Ai for everyone
 
Machine learning applications nurturing growth of various business domains
Machine learning applications nurturing growth of various business domainsMachine learning applications nurturing growth of various business domains
Machine learning applications nurturing growth of various business domains
 
A Practical Guide to AI and Automation
A Practical Guide to AI and AutomationA Practical Guide to AI and Automation
A Practical Guide to AI and Automation
 

Mehr von Tannistho Ghosh

How to use your customer facing educational materials on social media
How to use your customer facing educational materials on social mediaHow to use your customer facing educational materials on social media
How to use your customer facing educational materials on social mediaTannistho Ghosh
 
How can learning design support in D2C brand building
How can learning design support in D2C brand buildingHow can learning design support in D2C brand building
How can learning design support in D2C brand buildingTannistho Ghosh
 
Gamifying B2B sales
Gamifying B2B salesGamifying B2B sales
Gamifying B2B salesTannistho Ghosh
 
5 quick fix methods of transitioning your institute online manipal tech
5 quick fix methods of transitioning your institute online manipal tech5 quick fix methods of transitioning your institute online manipal tech
5 quick fix methods of transitioning your institute online manipal techTannistho Ghosh
 
A PRIMER ON DIGITAL MARKETING
A PRIMER ON DIGITAL MARKETINGA PRIMER ON DIGITAL MARKETING
A PRIMER ON DIGITAL MARKETINGTannistho Ghosh
 
Linkedin company page
Linkedin company pageLinkedin company page
Linkedin company pageTannistho Ghosh
 
Security threats in social networks
Security threats in social networksSecurity threats in social networks
Security threats in social networksTannistho Ghosh
 
A primer on virtual currency
A primer on virtual currencyA primer on virtual currency
A primer on virtual currencyTannistho Ghosh
 
A primer on ux design
A primer on ux designA primer on ux design
A primer on ux designTannistho Ghosh
 
High street labels strategy presentation
High street labels   strategy presentationHigh street labels   strategy presentation
High street labels strategy presentationTannistho Ghosh
 
How do you decode your brand's dna
How do you decode your brand's dnaHow do you decode your brand's dna
How do you decode your brand's dnaTannistho Ghosh
 
Some conversation starters for Rupa and Co.
Some conversation starters for Rupa and Co.Some conversation starters for Rupa and Co.
Some conversation starters for Rupa and Co.Tannistho Ghosh
 
5 #startups that can change your #day
5 #startups that can change your #day5 #startups that can change your #day
5 #startups that can change your #dayTannistho Ghosh
 
HOW TO CREATE GREAT BRAND SOLUTIONS!
HOW TO CREATE GREAT BRAND SOLUTIONS!HOW TO CREATE GREAT BRAND SOLUTIONS!
HOW TO CREATE GREAT BRAND SOLUTIONS!Tannistho Ghosh
 
Mystery of the Indian Consumer
Mystery of the Indian Consumer Mystery of the Indian Consumer
Mystery of the Indian Consumer Tannistho Ghosh
 

Mehr von Tannistho Ghosh (19)

How to use your customer facing educational materials on social media
How to use your customer facing educational materials on social mediaHow to use your customer facing educational materials on social media
How to use your customer facing educational materials on social media
 
How can learning design support in D2C brand building
How can learning design support in D2C brand buildingHow can learning design support in D2C brand building
How can learning design support in D2C brand building
 
Gamifying B2B sales
Gamifying B2B salesGamifying B2B sales
Gamifying B2B sales
 
5 quick fix methods of transitioning your institute online manipal tech
5 quick fix methods of transitioning your institute online manipal tech5 quick fix methods of transitioning your institute online manipal tech
5 quick fix methods of transitioning your institute online manipal tech
 
A PRIMER ON DIGITAL MARKETING
A PRIMER ON DIGITAL MARKETINGA PRIMER ON DIGITAL MARKETING
A PRIMER ON DIGITAL MARKETING
 
Linkedin company page
Linkedin company pageLinkedin company page
Linkedin company page
 
Uber for india
Uber for indiaUber for india
Uber for india
 
Security threats in social networks
Security threats in social networksSecurity threats in social networks
Security threats in social networks
 
A primer on virtual currency
A primer on virtual currencyA primer on virtual currency
A primer on virtual currency
 
Media ant [prez
Media ant [prezMedia ant [prez
Media ant [prez
 
A primer on ux design
A primer on ux designA primer on ux design
A primer on ux design
 
Uberizing Bengal
Uberizing Bengal Uberizing Bengal
Uberizing Bengal
 
The mobile store
The mobile storeThe mobile store
The mobile store
 
High street labels strategy presentation
High street labels   strategy presentationHigh street labels   strategy presentation
High street labels strategy presentation
 
How do you decode your brand's dna
How do you decode your brand's dnaHow do you decode your brand's dna
How do you decode your brand's dna
 
Some conversation starters for Rupa and Co.
Some conversation starters for Rupa and Co.Some conversation starters for Rupa and Co.
Some conversation starters for Rupa and Co.
 
5 #startups that can change your #day
5 #startups that can change your #day5 #startups that can change your #day
5 #startups that can change your #day
 
HOW TO CREATE GREAT BRAND SOLUTIONS!
HOW TO CREATE GREAT BRAND SOLUTIONS!HOW TO CREATE GREAT BRAND SOLUTIONS!
HOW TO CREATE GREAT BRAND SOLUTIONS!
 
Mystery of the Indian Consumer
Mystery of the Indian Consumer Mystery of the Indian Consumer
Mystery of the Indian Consumer
 

KĂŒrzlich hochgeladen

Apidays Singapore 2024 - Building Digital Trust in a Digital Economy by Veron...
Apidays Singapore 2024 - Building Digital Trust in a Digital Economy by Veron...Apidays Singapore 2024 - Building Digital Trust in a Digital Economy by Veron...
Apidays Singapore 2024 - Building Digital Trust in a Digital Economy by Veron...apidays
 
Slack Application Development 101 Slides
Slack Application Development 101 SlidesSlack Application Development 101 Slides
Slack Application Development 101 Slidespraypatel2
 
2024: Domino Containers - The Next Step. News from the Domino Container commu...
2024: Domino Containers - The Next Step. News from the Domino Container commu...2024: Domino Containers - The Next Step. News from the Domino Container commu...
2024: Domino Containers - The Next Step. News from the Domino Container commu...Martijn de Jong
 
Handwritten Text Recognition for manuscripts and early printed texts
Handwritten Text Recognition for manuscripts and early printed textsHandwritten Text Recognition for manuscripts and early printed texts
Handwritten Text Recognition for manuscripts and early printed textsMaria Levchenko
 
A Year of the Servo Reboot: Where Are We Now?
A Year of the Servo Reboot: Where Are We Now?A Year of the Servo Reboot: Where Are We Now?
A Year of the Servo Reboot: Where Are We Now?Igalia
 
Driving Behavioral Change for Information Management through Data-Driven Gree...
Driving Behavioral Change for Information Management through Data-Driven Gree...Driving Behavioral Change for Information Management through Data-Driven Gree...
Driving Behavioral Change for Information Management through Data-Driven Gree...Enterprise Knowledge
 
The Role of Taxonomy and Ontology in Semantic Layers - Heather Hedden.pdf
The Role of Taxonomy and Ontology in Semantic Layers - Heather Hedden.pdfThe Role of Taxonomy and Ontology in Semantic Layers - Heather Hedden.pdf
The Role of Taxonomy and Ontology in Semantic Layers - Heather Hedden.pdfEnterprise Knowledge
 
The Codex of Business Writing Software for Real-World Solutions 2.pptx
The Codex of Business Writing Software for Real-World Solutions 2.pptxThe Codex of Business Writing Software for Real-World Solutions 2.pptx
The Codex of Business Writing Software for Real-World Solutions 2.pptxMalak Abu Hammad
 
Boost Fertility New Invention Ups Success Rates.pdf
Boost Fertility New Invention Ups Success Rates.pdfBoost Fertility New Invention Ups Success Rates.pdf
Boost Fertility New Invention Ups Success Rates.pdfsudhanshuwaghmare1
 
Axa Assurance Maroc - Insurer Innovation Award 2024
Axa Assurance Maroc - Insurer Innovation Award 2024Axa Assurance Maroc - Insurer Innovation Award 2024
Axa Assurance Maroc - Insurer Innovation Award 2024The Digital Insurer
 
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...Drew Madelung
 
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
08448380779 Call Girls In Diplomatic Enclave Women Seeking MenDelhi Call girls
 
CNv6 Instructor Chapter 6 Quality of Service
CNv6 Instructor Chapter 6 Quality of ServiceCNv6 Instructor Chapter 6 Quality of Service
CNv6 Instructor Chapter 6 Quality of Servicegiselly40
 
04-2024-HHUG-Sales-and-Marketing-Alignment.pptx
04-2024-HHUG-Sales-and-Marketing-Alignment.pptx04-2024-HHUG-Sales-and-Marketing-Alignment.pptx
04-2024-HHUG-Sales-and-Marketing-Alignment.pptxHampshireHUG
 
08448380779 Call Girls In Friends Colony Women Seeking Men
08448380779 Call Girls In Friends Colony Women Seeking Men08448380779 Call Girls In Friends Colony Women Seeking Men
08448380779 Call Girls In Friends Colony Women Seeking MenDelhi Call girls
 
🐬 The future of MySQL is Postgres 🐘
🐬  The future of MySQL is Postgres   🐘🐬  The future of MySQL is Postgres   🐘
🐬 The future of MySQL is Postgres 🐘RTylerCroy
 
TrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
TrustArc Webinar - Stay Ahead of US State Data Privacy Law DevelopmentsTrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
TrustArc Webinar - Stay Ahead of US State Data Privacy Law DevelopmentsTrustArc
 
A Domino Admins Adventures (Engage 2024)
A Domino Admins Adventures (Engage 2024)A Domino Admins Adventures (Engage 2024)
A Domino Admins Adventures (Engage 2024)Gabriella Davis
 
Artificial Intelligence: Facts and Myths
Artificial Intelligence: Facts and MythsArtificial Intelligence: Facts and Myths
Artificial Intelligence: Facts and MythsJoaquim Jorge
 
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024The Digital Insurer
 

KĂŒrzlich hochgeladen (20)

Apidays Singapore 2024 - Building Digital Trust in a Digital Economy by Veron...
Apidays Singapore 2024 - Building Digital Trust in a Digital Economy by Veron...Apidays Singapore 2024 - Building Digital Trust in a Digital Economy by Veron...
Apidays Singapore 2024 - Building Digital Trust in a Digital Economy by Veron...
 
Slack Application Development 101 Slides
Slack Application Development 101 SlidesSlack Application Development 101 Slides
Slack Application Development 101 Slides
 
2024: Domino Containers - The Next Step. News from the Domino Container commu...
2024: Domino Containers - The Next Step. News from the Domino Container commu...2024: Domino Containers - The Next Step. News from the Domino Container commu...
2024: Domino Containers - The Next Step. News from the Domino Container commu...
 
Handwritten Text Recognition for manuscripts and early printed texts
Handwritten Text Recognition for manuscripts and early printed textsHandwritten Text Recognition for manuscripts and early printed texts
Handwritten Text Recognition for manuscripts and early printed texts
 
A Year of the Servo Reboot: Where Are We Now?
A Year of the Servo Reboot: Where Are We Now?A Year of the Servo Reboot: Where Are We Now?
A Year of the Servo Reboot: Where Are We Now?
 
Driving Behavioral Change for Information Management through Data-Driven Gree...
Driving Behavioral Change for Information Management through Data-Driven Gree...Driving Behavioral Change for Information Management through Data-Driven Gree...
Driving Behavioral Change for Information Management through Data-Driven Gree...
 
The Role of Taxonomy and Ontology in Semantic Layers - Heather Hedden.pdf
The Role of Taxonomy and Ontology in Semantic Layers - Heather Hedden.pdfThe Role of Taxonomy and Ontology in Semantic Layers - Heather Hedden.pdf
The Role of Taxonomy and Ontology in Semantic Layers - Heather Hedden.pdf
 
The Codex of Business Writing Software for Real-World Solutions 2.pptx
The Codex of Business Writing Software for Real-World Solutions 2.pptxThe Codex of Business Writing Software for Real-World Solutions 2.pptx
The Codex of Business Writing Software for Real-World Solutions 2.pptx
 
Boost Fertility New Invention Ups Success Rates.pdf
Boost Fertility New Invention Ups Success Rates.pdfBoost Fertility New Invention Ups Success Rates.pdf
Boost Fertility New Invention Ups Success Rates.pdf
 
Axa Assurance Maroc - Insurer Innovation Award 2024
Axa Assurance Maroc - Insurer Innovation Award 2024Axa Assurance Maroc - Insurer Innovation Award 2024
Axa Assurance Maroc - Insurer Innovation Award 2024
 
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...
 
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
 
CNv6 Instructor Chapter 6 Quality of Service
CNv6 Instructor Chapter 6 Quality of ServiceCNv6 Instructor Chapter 6 Quality of Service
CNv6 Instructor Chapter 6 Quality of Service
 
04-2024-HHUG-Sales-and-Marketing-Alignment.pptx
04-2024-HHUG-Sales-and-Marketing-Alignment.pptx04-2024-HHUG-Sales-and-Marketing-Alignment.pptx
04-2024-HHUG-Sales-and-Marketing-Alignment.pptx
 
08448380779 Call Girls In Friends Colony Women Seeking Men
08448380779 Call Girls In Friends Colony Women Seeking Men08448380779 Call Girls In Friends Colony Women Seeking Men
08448380779 Call Girls In Friends Colony Women Seeking Men
 
🐬 The future of MySQL is Postgres 🐘
🐬  The future of MySQL is Postgres   🐘🐬  The future of MySQL is Postgres   🐘
🐬 The future of MySQL is Postgres 🐘
 
TrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
TrustArc Webinar - Stay Ahead of US State Data Privacy Law DevelopmentsTrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
TrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
 
A Domino Admins Adventures (Engage 2024)
A Domino Admins Adventures (Engage 2024)A Domino Admins Adventures (Engage 2024)
A Domino Admins Adventures (Engage 2024)
 
Artificial Intelligence: Facts and Myths
Artificial Intelligence: Facts and MythsArtificial Intelligence: Facts and Myths
Artificial Intelligence: Facts and Myths
 
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
 

Simplifying ai: What to use when?

  • 2. We read about the use ML models and ML Ops, Deep Learning and even sentient AI everyday...
  • 3. We get the shivers when we see machines behaving like impulsive people on Netflix movies...
  • 4. How much do we know about them? Do we know what they mean? And what they mean for our business?
  • 5. Let's Uncomplicate by keeping Netflix aside... 2020
  • 6. Let's Uncomplicate by keeping Netflix aside... 2020
  • 7. Let's define the coolest term first WHAT IS AI?
  • 8. When a machine starts to perform cognitive functions that we associate with human minds such as perceiving, reasoning, learning and problem solving - we call it 'Artificial Intelligence'.
  • 9. The term was coined by Prof John McCarthy which he explained as 'the science and engineering of making intelligent machines'.
  • 10. Computer Vision, Robotics, Natural Language Processing, Machine Learning are all technologies that enable AI.
  • 11. At the heart of AI is machine learning 2020
  • 12. What does Machine Learning do? Most recent advances in AI has been done through Machine Learning. ML Algorithms detect patterns in large sets and learn to make predictions and recommendations on their own.
  • 13. The coolest part is - the ML algorithms can adapt and improve themselves over time as more data becomes accessible. 2020
  • 14. Types of ML Outputs Descriptive Describe what happened. This is heavily used across industries. E.g - LMS analytics, learner behavior. 01. Prescriptive Describe what will happen. This is inherently probabilistic. e.g: Financial Monitoring. 02. Describe What to do. e.g: Making investment predictions or predicting learner drop-out rates. Predictive 03. There are three types of analytics. Out of them Prescriptive and Predictive are key domains of ML outputs. MACHINE LEARNING
  • 15. The three types of Machine Learning: Type 1 Supervised Learning We give the input data We give the output variables We label all the data We train the algorithm to make connections between the input and output variables. Then we have coffee as the algorithm works on its own
  • 16. The three types of Machine Learning: Type 2 Unsupervised Learning We give the input data The algorithm figures out the output variable on its own. E.g: An algorithm figures out types of probable customers for a product from social media chatter. We have coffee after giving only the input data
  • 17. The three types of Machine Learning: Type 3 Reinforcement Learning An algorithm figures out how to do a particular task by getting rewarded for doing it right and punished for doing it wrong. E.g: If a self-driving car hits a pedestrian you take away its fuel. That's punishment. We stand behind the algorithm with a stick here.
  • 18. And there's Transfer Learning too... When a ML algorithm has worked something out in a particular domain, it is easier to apply the same knowledge and learn something faster in a adjacent areas.
  • 19. How do we use these algos in actual business cases 2020
  • 20. Let's start with the Supervised Learning Algorithms 2020
  • 21. Linear Regression THE A-Z OF MAKING IT STICK 2020 Linear Regression is a standard method for modeling the past relationship between independent variables and dependent variables to help predict future values of the output. This can help you understand product-sales drivers e.g - prices of competition, distribution, etc. This can help you optimize price points and estimate product-price elasticities. What is it? Where to use?
  • 22. Logistic Regression THE A-Z OF MAKING IT STICK 2020 An extension of Linear Regression -Logistic regression is used for classification tasks where the output is always in a binary (a Yes or No answer). Logistic regression is used in areas where the output variable needs to be a yes/no answer, for example: will someone pay off a loan, predict if a skin lesion is malignant or benign. What is it? Where to use?
  • 23. Linear/Quadratic Discriminant Analysis THE A-Z OF MAKING IT STICK 2020 A Linear Discriminant Analysis upgrades a logistic regression to deal with non-linear problems so that if you make changes in the input variables they do not have proportional changes to the output variables. If you are a SaaS firm, this can help you predict the possibility of Client churn, or even the potential of a sales person closing a lead. What is it? Where to use?
  • 24. Decision Trees THE A-Z OF MAKING IT STICK 2020 A highly interpretable classification or regression model that splits data-features into branches at decision nodes until a final decision output is made. If you are building a product, Decision Trees can help you identify which feature is most likely to drive sales/adoption of the product. What is it? Where to use?
  • 25. Naive Bayes THE A-Z OF MAKING IT STICK 2020 Classification technique that applies Bayes theorem which allows the probability of an event to be calculated based on the knowledge of factors that might affect it. You can use Naive Bayes to analyze sentiments - such as understanding your product perception in the market or even create classifiers to identify to filter spam emails (Naive Bayes can help you classify emails which say 'I am a sad billionnaire in Nigeria' and automatically put them in spam. What is it? Where to use?
  • 26. Naive Bayes THE A-Z OF MAKING IT STICK 2020 Classification technique that applies Bayes theorem which allows the probability of an event to be calculated based on the knowledge of factors that might affect it. You can use Naive Bayes to analyze sentiments - such as understanding your product perception in the market or even create classifiers to identify to filter spam emails (Naive Bayes can help you classify emails which say 'I am a sad billionnaire in Nigeria' and automatically put them in spam. What is it? Where to use?
  • 27. Support Vector Machine THE A-Z OF MAKING IT STICK 2020 Support Vector Machine is a technique that's typically used for classification but can be transformed to perform regression. It draws an optimal division between classes You can predict how many patients a hospital will need to serve in a given time period. Support Vector Machine algorithms are also applied in text classification, image recognition, face recognition, etc. What is it? Where to use?
  • 28. Random Forest THE A-Z OF MAKING IT STICK 2020 Random Forest improves the accuracy of a single Decision Tree by generating multiple decision trees and then identifying the best output by taking a majority vote. You can predict call volumes in call centers and take relevant hiring and training decisions by using the Random Forest algorithm. What is it? Where to use?
  • 29. AdaBoost THE A-Z OF MAKING IT STICK 2020 Think of an AdaBoost (or Adaptive Boosting) is iike a discerning CEO who troubles everyone for the best result. The AdaBoost is a classification technique that runs a multitude of models to come up with a decision and weighs them based on their accuracy. You can predict call volumes in call centers and take relevant hiring and training decisions by using the Random Forest algorithm. What is it? Where to use?
  • 30. Gradient Boosting Trees THE A-Z OF MAKING IT STICK 2020 A classification technique where the decision trees come in a sequential manner by correcting the errors of the previous decision trees. The final output is a combination of the results of all the trees. This can be used in forecasting - such as forecasting of demand or inventory levels in eCommerce. What is it? Where to use?
  • 31. Simple Neural Network THE A-Z OF MAKING IT STICK 2020 A model in which artificial neurons make up three layers ( an input layer, a hidden layer and an output layer) that can be used to classify data or find relationships between variables in regression problems. Predict the probability of a user paying a certain price for a particular service Can be used in SaaS services, OTT services, healthcare programs, etc. What is it? Where to use?
  • 32. Whew! Wasn't that quite a bit! Here are some quick quips on where supervised learning is going!
  • 33. For supervised learning to run effectively - it needs tons of labeled data WHICH, FRANKLY... IS A CHALLENGE.
  • 34. That's why Tesla is using SELF SUPERVISED LEARNING. THE A-Z OF MAKING IT STICK 2020
  • 35. Supervised learning can be time consuming and may take months to create the right dataset to train the ML model. SELF-SUPERVISED LEARNING IS A MACHINE LEARNING APPROACH WHERE THE MODEL TRAINS ITSELF BY LEVERAGING ONE PART OF THE DATA TO PREDICT THE OTHER PART AND GENERATE LABELS ACCURATELY.
  • 36. Robotic surgeries: SSL can help in robotic surgeries by estimating the depth of the human body and providing better medical visuals. Autonomous cars: SSL can estimate the roughness of the terrain. Where can Self Supervised Learning be effective:
  • 37. Let's move on to Unsupervised Learning
  • 38. K-Means Clustering The algorithm helps put data into a number of groups (K) each of which contain data of similar characteristics. K-Means Clustering can help you identify customer types, customer preferences, and even customer personas from unstructured data such as social media conversations. What is it? Where to use?
  • 39. K-Means Clustering The algorithm helps put data into a number of groups (K) each of which contain data of similar characteristics. K-Means Clustering can help you identify customer types, customer preferences, and even customer personas from unstructured data such as social media conversations. What is it? Where to use?
  • 40. Gaussian Mixture Model A generalized version of K-Means Clustering, the GMM allows for more flexibility in the size and shape of the clusters Gaussian Mixture Model can help in better segmentation of customers using less-distinct customer characteristics. For example, you can use the GMM to identify the chances of employee attrition. What is it? Where to use?
  • 41. Hierarchical Clustering It aggregates clusters in a hierarchical tree to form a classification system. You can use the Hierarchical Clustering algorithm to divide your loyalty program customers into micro-segments to devise a more personalized approach towards customer-outreach and increase ARR. What is it? Where to use?
  • 42. Recommender Systems A recommender system often uses cluster behavior prediction to identify the important data necessary for making a recommendation We did get back to Netflix didn't we, even though we kept it aside. Netflix uses the Recommender system to make personalized movie recommendations to you. Next time you get a Korean Sci Fi recommendation after watching a Russian Sci Fi, just remember what was being used. What is it? Where to use?
  • 43. Now for mistakes and punishment s.... You got it! We are talking about Reinforcement learning here.
  • 44. Some Use Cases here Reinforcement learning can be used to make the right trade in time. The way you train your algo is by punishing it every-time it makes the wrong trade. Using a multi-modal input model you train your self-driving car to avoid mistakes, potholes, hitting pedestrians on the road, going in the wrong road, etc. Over a period of time through a cycle of rewards and punishments your algo optimizes its driving behavior. Algorthmic trading Optimizing driving behaviour of self- driving cars
  • 45. And now for Deep Learning It's the key tech behind some fo the coolest inventions such as Self- Driving cars, voice control behind phones, etc.
  • 47. Deep Learning is a subset of Machine Learning that uses algorithms inspired by the structure and functions of the brain's neural network
  • 48. Deep Learning and Artificial Neurons receive input values, weighs these values with weights and returns an output value. Deep Learning works with artificial neurons. "Artificial Neurons" are mathematical tools that The value returned by the neuron is called an 'activation' similar to the electrical signals that a human neuron receives from its dendrites and decides if it needs to activate it to other neurons.
  • 49. How does Deep Learning work? In Deep Learning, an Artificial Neural Network passes information through a series of neurons in an input layer. Think of your sensory organs (ear, nose) taking in information and passing onto your brain. The hidden layers at the second level processes this information and assigns weights to it. For example, if you are trying to predict an inventory - the weightage given to past sales history will be higher than the week when it has been sold. Deep Learning is used in predicting and completing complex tasks such as predicting the weather.
  • 50. Types of Deep Learning Neural Networks Recurrent Neural Networks Convolutional Neural Networks Multilayered Perceptrons The three popular types of Deep Learning Neural Networks include the following:
  • 51. Convolutional Neural Networks Recurrent Neural Networks A ConvNet (employed in areas like Computer Vision) use a range of convolutional operations to extract the features given an image or video. A ConvNet or CNN is a favourite when it comes to areas like image classification, image extraction, videos, etc. The applications of CNN can be seen in a range of industries starting from surveillence, creative, medicine, etc. RNN uses sequential data feeding. Thus, the inputs in RNN include current input as well as previous samples. RNN is good with time series calculations and handling sequential data. The application of RNN include text predictions - if you are about to type something in the search box and you get a series of choices based on your previous inputs as well as common inputs of others, then it is RNN at play. The applications of RNN go into understanding human language and has implications in Natural Language Processing and Natural Language Understanding.
  • 52. Multilayer Perceptrons Multilayer perceptron is a classical neural network used for basic operations like data visualization, data compression, and encryption. It basically has a input layer, hidden layers and output layer Multilayer Perceptrons are used in areas like data visualization, data analysis, etc.
  • 53. WHOA! WASN'T THAT A LOT? BUT BEFORE WE GO... HERE'S A SHORT TEASER...
  • 54. Sunspring is a 2016 movie written by an AI called Benjamin. Can you guess the tech that was used in training up the AI here?