HOW DO THEY DO IT | Top 10 Applications of Machine Learning | Intellipaat

HOW DO THEY DO IT | Top 10 Applications of Machine Learning | Intellipaat


Have you ever thought how Google map predicts the traffic
so accurately or how Amazon is recommending you the products or even how the self-driving cars work if
Yes, then you have landed yourself at just the right place
Hello everyone I am Atul from Intellipaat and today I’m here with a list of top 10
Applications of machine learning and how do they do it?
I’d suggest you to stay tuned till the end where I’ll show you how we have already stepped into the future world with machines
So starting with on number 10, we have as google map traffic prediction
Well, google map is very accurate at predicting traffic. But how Google map predicts the traffic. How does it get this information?
Well, if you have an Android phone or an iPhone with Google map open and services enabled on it then your phone or the app
Anonymously sends real-time data back to the Google
The company then uses this information or the data to calculate. How many cars are there on the road or
How fast they are moving. So more people using the app implies more accurate as the traffic data
Google has also incorporated traffic data from app called waze
The company bought it for 1 billion dollar in 2013
the app monitors traffic report from the local Department of Transportation
Google even keeps a history of traffic patterns on specific role so that it can predict
What would be the traffic at a specific time on a specific location?
If the traffic is more the app would suggest you a faster route to reach your destination and time
thereby making you build trust in the app.
If you’re feeling like Google is monitoring and tracking your every move then don’t worry
You can still opt out at any time just by turning off the location services
But what if everyone did that?
Well, we would be in a trouble as the data and the result
Might not be as accurate. There’s the reason why at some places Google map is not so accurate
So this was about Google map using machine learning algorithm to analyze and predict the result from your data
So more the data you feed to the app the more accurate it gets
All right, so moving on ahead
Next on number 9 we have is google translate
Well, Google Translate is a free tool that enables you to translate sentences
Documents and even whole websites instantly, but how exactly does it work
How google translator uses machine learning to translate your text
Well, it must seem like Google has a room full of bilingual Elves working for them
But the matter of factors all of the translations come from the computers
These computers uses a process called statistical machine translation
Which is just a fancy way to say that the computers generate translation based on patterns found in large amount of text
But let’s take a step back if you want to teach someone a new language. How would you start?
Well, you might start by teaching them vocabulary words romantical rule and then explain how to construct sentences
Right!
Similarly when a computer learns a foreign language it understands it in the same way
That is by referring to the vocabulary and the set of rules
but if you see
Learning a new language is very complicated and any language learner can tell you that there are exceptions to almost any rule
When you try to capture all of these exceptions and exceptions to the exceptions in a computer program
The translation quality begins to break down
Now when it comes to Google Translate take a slightly different approach for it
So instead of trying to teach computers all the rules for a language what it does it lets the computer discover the rules for themselves
again, one of the tasks of machine learning
Well, they do this by analyzing millions and millions of documents that have been translated by human translators
These translated texts come from books
Organizations like UN and websites from all around the world
The computer scan these decks looking for statistically significant patterns
That is to say pattern between the translation and the original text that are unlikely to occur by chance
Once the computer finds a pattern you can use this pattern to translate similar texts in future
Now when you repeat this process billions of time you end up with billions of patterns and one very smart come a program
For some languages. Hi were the google translator have fewer translated documents available and
therefore fewer patents to be detected by this software
well, this is the reason why our
translation quality differs by language and language fair
The Google’s translation aren’t always perfect
But by constantly providing new translated text we can make our computer smarter and our translation better
so next time when you translate a sentence or webpage with Google Translate think about those millions of documents and billions of pattern that
Ultimately led to the translation of all of it happening in a blink of a eye using the machine learning algorithms
Pretty cool, isn’t it? So this was about how Google translates your speech?
So moving on ahead
Next on number eight we have is auto oil text by Facebook of
Facebook’s automatic oil text so what is that?
Let’s see. Well nowadays
Facebook has rolled out a new feature that lets the blind users
Extort the internet and see pictures that their friends are posted like never before
This feature is called automatic alternative text
It’s just one of the way that companies are changing the way blind people are experiencing the world on and offline
well, if you see according to World Health Organization, there are more than 39 million blind people around the globe and
Just because they cannot see it doesn’t mean that they cannot use the Internet
But the way they experience it is very different from how we cite that people do
Wait talking about you. You can see a whole screen full of information and
Because of that you are able to make decisions about how to get the information you need
But for a blind person it’s pet more awkward
The blind relies on the two will call screen readers to narrate websites or apps or translate them into Braille
That is one line of text or link at a time
so instead of using the mouse, they generally jump around the page using keyboard shortcuts and
Because this screen reader works by looking at the website code not just what is visible on a display
Well, sometimes the site can be almost incomprehensible if they are not designed to be read by the tech
The early days of the internet were actually easier for the blind user because almost everything was text but over time not just content
But also the design has become much more visual in nature
Social media has become a big part of this shift
Facebook says that people share more than two billions of 400 days across its various products
But those picture would be of no use for the blind if they don’t come with the text that describe what they show right?
So Facebook is trying to fix this problem with automatic alternative text
It works like when a user has built-in reader turn on and select a picture
they are then Facebook uses artificial intelligence algorithms to detect basic features in the image and create a new alt text that
The screen reader will share with the user basically
Describing what’s there in the photo?
Just for example
If the picture shows a couple overlooking the ocean while wearing sunglasses the new alt text would be something like this
The image may contain two people smiling
sunglasses sky out of water
It’s not exactly poetry eventually
They’re hoping to make the alt text more like a narrative but it still give a blind user new way to imagine. What’s the animage?
Well Facebook is not the only one who was trying to make internet more usable for the blind people
Twitter has also recently announced a feature that will finally allow the users to attach an oiled image or the oil
Text to the images they have uploaded
Fine. So this was about how Facebook is using machine learning and artificial intelligence to help the blind
Next on number seven we have as Amazon’s recommendation engine
So, how do you think Amazon recommends your next product?
Well, the answer is machine learning with big data. So how they are helping in recommendation system engine of Amazon
Well, it involves three stages
events ratings and filtering in the even phase
Amazon tracks and stores data on all customer behavior and activity on the site
Every click you make is an even and the record of the user has logged into the database
the entry is stored as something like user a click X product once
It’s like user a click
extra duck once so different kind of evens are captured for different kinds of actions like
User liking a product adding product to the cart or purchasing a product fine
Next thing comes is the rating. Well ratings are important as they reveal what a user feel about the product
Recommendation system then assign implicit values on different kinds of user actions like for star for purchase
Three staff alike and two stuff click and so on now this recommendation system
It also uses NLP or the national language processing to analyze the feedback which is provided by the user
like a feedback can be
The product was great. But packaging was not good at all
So using natural language processing the recommendation system calculate the sentiment score and classifies the feedback as positive negative or neutral
fine
Till here it was about reading so alas and the third stages filtering
Well in this step machine filter the product based on rating and other user data
recommendation system uses different kinds of filtering like collaborative filtering user base filtering and hybrid filtering
Collaborating filtering is the one in which all the users choices are compared and they get a recommendation
for example user X likes a product a B C and D and
Use a vite likes the product a B C D and E
So there’s a chance that user X will also like the product e and the Machine will recommend product e to the user X
fine
Next comes the user base filtering well in this the users browsing history likes
Purchases and ratings are taken into account
before providing the recommendation and
finally the next SP hybrid approach
Well, there’s a mix and match of both the collaborative and user base type of filtering. So this is how Amazon recommends your next product
Well, this recommendation is not just made in Amazon. But even in Alibaba eBay and Flipkart to using the same approach
fine
So moving on ahead on number six we have spam detection by Gmail
So, how do you think Gmail understands which email is spam or not? Let’s see
Well spam detection basically works on the base of filter
The settings that are constantly updated based on new technologies new spam
identifications and the feedback given by Gmail user about potential spammers
The spam filters use either the text filters or it eliminates the threads based on senders and their history
Like whether the sender was reported or not fine
So first we have is a text filter
So X filters work by using algorithms to detect which word and phrase are most often used in spam email
Phrases like law tree or you on a free bitcoin are often an immediate flag for removal by filters
Stomata have gotten voiced these days
So they often use blunt misspellings or even substitute of characters like three dollars in order to make it pass through the filters
Fortunately if the modern spam filters can also make allowance for these types of misspellings
And if a word comes up even with a character substitution, the message will still be blocked
All right
next type of filter that we have is client filters in
Addition to the simple text filter algorithm the top spam filters uses client identity and history to block
malicious and annoying spam emails
Now, how does the spam filter work? Well, this is done by looking at all of the messages a user has sent out if
this sends a huge amount of emails constantly or if
Several of their messages have already being marked as a spam through the text filter their emails will often be blocked entirely
So this bring us to the use of blacklist?
spam filters also include a blacklist
Blacklisting is simply a process of adding the known email addresses of a spammer to the list
This list prevents any inbound messages from the email addresses going forward if you have marked something as a spam
it will directly go to this blacklist and it may also help the Google or the Gmail to make a note of certain keywords from
That spam mail. Well, that’s the outline of spam detection or how Gmail understands which email is spam
But the real time process a lot more complex and it consumes a lot of data to fine
So let’s move ahead
So number five we have is Amazon Alexa
So how Amazon Alexa understands your command?
well the device which you are seeing on your screen is Amazon echo and the brain or the voice of echo is known as
Amazon Alexa
Will the device has enough built-in smarts to do a number of tasks like playing back music and making lights blink?
The device has enough built-in smarts to do a number of tasks
Like playing back music giving the weather report or even making the lights blink
It can also recognize the elixir name when you say the word elixir
recognises the void
Amazon calls this as a weak word and
When you say the word Alexa it starts recording your voice
So when you are finished speaking, it sends this recording over the Internet to Amazon
The service that process this recording is called Alexa voice services or AVS
The service is run by Amazon it converts the recording into commands that it interprets
It’s more than a simple voice to text hours
It’s a fully programmable service that can work with other online services to do a surprising range of things
Once authorized by Amazon anyone can use the service for free to build a homemade echo
Well, amazon offers sample code for building one using a Raspberry Pi a simple $30 computer
Well, it might sound extremely selfless of Amazon to provide the service of free but as always they have their reasons
Amazon wants other to build the service into their product so they can sell you these stuff
So every product that has alexa built in is a device that can be used to buy stuff from Amazon
These commands that Alexa interprets can be very simple
If you ask for time the Avs sends back an audio file of Alexa telling you the time
Which the echo plays back they can also be more complex
Like if you tell Alexa to play Pink Floyd the Avs will search the music service
You have set up for Pink Floyd then send a command back to the echo that play the requested music
Alexa can also work with other technologies in a home and beyond for example
If you have set up any Philips hue smart bulbs
It can control these you can ask Alexa to turn on the living room lights
And Alexa will send a command to the echo that sends a command to these light bulbs to turn on
It can also work at online services
You can link elixir to uber and you can request overdrive simply by asking Alexa
You can even link it to Domino’s and you can order pizza with your voice
Well this approach means that echo and Alexa can do a lot of things and the list is just getting longer
Amazon is adding more feature called skills to Alexa
And if you are smart enough, you can build the skills on your own
This means that you can even use Alexa to control things that are not on the supported list
the hackers have been working hard to make a do things like
Adding support for controlling the media center program Kodi and figuring out when the next bus will arrive at your local bus stop
How this approach is also called the Achilles heel of Alexa
it needs internet connectivity and AWS to work if your connection is slow or isn’t working nxr won’t be available and
Your eco will be useless
Well in future if Amazon decides you to charge for the service or just close it down
You will be left with the useless device
When if you consider these products Amazon is not the only company in the market Google Apple and Microsoft
Offer services that can perform tasks by voice command in the form of ok, Google Siri and Cortana
Even they are using the same approach that as voice commands that are processed in a cloud service
But most of them are not as flexible or as integrated with service as Alexa’s
So whichever one of these services end up being the one we all use hopefully they would be as polite as Alexa
Well, I will just give you an example
When I ask nxr how she works
She replied lots of people have worked hard to teach me and I’m still learning
Wouldn’t it be nice if Ola up the answers with dart modest and polite right? So let’s move ahead
So next on number four we have as Tesla’s self-driving car
Well a recent study has shown that who were 90% of road accidents are caused by human error
To err is human
But behind the wheel mistakes are often catastrophic the accidents have led to a massive number of unnecessary deaths
Lies that could have been saved if they were driving safer. So this is where self-driving cars comes into picture
They are autonomous car which are much safer than human driven cars
They are unaffected by factors like driver fatigue emotions or illness
This makes them very safe
Self-driving cars are always at enter an active observing the environments and scanning multiple directions
It would be difficult to make a move that the car has not anticipated
So, isn’t that cool?
But the main question arises how does a self-driving car work?
Well these self-driving cars mainly use three different
Technologies the ID sensors the IOT connectivity and these software algorithms
So if you talk about IOT sensors, there are many types of sensors available today that of making autonomous cars a reality
sensors for blind spot monitoring forward collision warning
radar camera lidar and
Ultrasonic all work together to make the navigation of a self-driving car possible
Next is the IOT connectivity
self-driving cars uses cloud computing to act upon traffic data weather maps adjacent cars and
surface conditions among others
This help them to monitor their surroundings better and make informed decisions
Well self-driving cars must be connected to the Internet all the time
Even if edge computing Hardware can solve small computing tasks locally
So finally we have the software algorithm which made them work
So all the data the car collects need to be analyzed to determine the best course of action
Well, this is the main function of the control algorithm and software and this is where machine learning comes into picture
this is the most complex part of the cell driving car since it has to make decision flawlessly and correctly a
Flaw like an uber cell driving accident can be a fatal in
Today’s world the most famous self-driving cars are those made by Tesla and Google
Tesla cars work by analyzing the environment using a software system known as autopilot
These autopilot uses high-tech cameras to view and collect data from the world much like we use our I to do
using what’s called computer vision or sophisticated image recognition
It then interprets this information and makes the best decision based on it
has started earlier
Tesla’s self-driving car technology is already being sold in the market today. So
This was about how does self-driving car walk using the technology of machine learning?
Let me just show you a video this will make things more clear to you
So morning on ahead next on number three we have as Netflix movie recommendation
So more than 80% of the TV shows that you guys are watching on Netflix are discovered through the platforms recommendation system. That means
Majority of what you decide to watch on Netflix is a result of the decisions made by mysterious black box algorithm
More than 80% of the TV shows that you guys watch on Netflix are discovered through the platforms recommendation system
That means the majority of what you decide to watch on Netflix
As the result of the decisions made by a mysterious black box of an algorithm
So I interest in learning how it works
How the Netflix recommends you the movie?
well
Netflix uses machine learning algorithms to recommend you the list of movies and find shows that you might have not initially chosen
Well to do this, it looks at threads within the content rather than relying on the board journal to make its predictions
this explains how for example
One an eight people who watch Netflix Marvel Show are completely new to comic book based stuff on Netflix
Todd Yellin
The Netflix vice-president of product innovation has explained the working of Netflix as a three legged stool
The three legs are the stool would be Netflix members
second daggers who understand everything about the content and
The third big machine learning algorithms that take all of the data and put things together
Netflix has more than 100 million worldwide users
If you count the multiple user profiles for each subscribers, it comes to a total of around 250 million active profiles
Netflix uses different kind of data from these profiles like it keeps a track of what you guys are watching from your profile
What do you after completing your current video? And even what do you have watched earlier?
It also keeps a track of what you have watched a year ago or what you are currently watching
and at what time of the day so
This data it has the first leg of the metaphorical stool
Now they combine this information with more data to understand the content of the shows that you are watching
This data is gathered from dozens of in-house and freelance staff watch every minute or every show on Netflix and target
the tags
they use ranged from how cerebral the pieces to whether it has an assemble case or as set in space or
Soccer up cough all these tags and user behavior data are taken and as fee to a very sophisticated machine learning algorithm that figures out
What’s most important or what should it be? Like how much should it matter if a consumer has watched something yesterday?
Should that count twice as much but 10 times as much compared to what they have watched a whole year ago
How about a month ago?
How about if they watched ten minutes of the content and abandoned it or they blinked through it in two nights?
So how do Netflix pay all that?
Well, this is where machine learning comes in
Well what those three things created was taste communities around the world
It’s about people who watch the same kind of things that you watch
Viewers are made to fit into thousands of multiple taste groups
And it’s these that affect what recommendation pop up onto the top of your on-screen interface
With jar rows are displayed and how each row is ordered from each individual viewer
the tags that are used for machine learning
Algorithms are the same across the globe how a smallest subset of tags are used in a more outward facing wave
Feeding directly into the user interface and deferring depending on a country language and cultural context
The tabs have to be localized in a way that made sense. For example the word gritty as
In 3d drama may not translate into Spanish or French
Okay, the data that Netflix feeds into its algorithm can be broken down into two types implicit and explicit
Explicit data is what you literally tell us. For example, you give a thumbs up to the friends and the Netflix get it
Next is the implicit data
Implicit data is really behavioral data. It’s like you didn’t explicitly tell Netflix that you like black mirror
But you just blend on it and watched it in two nights
So here the Netflix understand that behaviorally well just as a matter of fact
Majority of useful data is implicit data. So this was about how Netflix is using machine learning to recommend you the movies
So moving on ahead next on number two. We have Keuka robots
Future is here already is it?
Well, Keuka is a leader manufacturer of industrial robots and solutions for factory automations
Their robots are manufactured to an advanced level of Robotics capable of performing tedious tasks
They are used in large companies predominantly in the automotive industry, but also in other industries such as the aerospace industry
Let me show you a short video on Keuka versus tim ball. The number one table tennis player in the world
This will show you how
accurate and precise the sku car robot can be
Fine in the end number one. We have mollies body kitchen
Have you ever thought how it would be like to tell a machine what you want to eat?
And the Machine cooks the exact same recipe and the exact same place for you
Well, if you think that this is the future then let me tell you guys a company called male have already created such a robot
The robot is capable enough to learn n number of dishes for you
Well, let me again show you a short video on it. This will make things more clear to you
Imagine having a five-star chef cooking all your meals for you every day of your life
This is the dream and the goal that moley robotics is trying to accomplish
They have created the first fully automated and intelligent robots that can learn recipes
Cook all kinds of foods with a remarkable precision and then clean up after themselves
moley robotics built a working prototype of the molle kitchen this
Futuristic looking workstation is equipped with advanced robotic arms and hands that can grip kitchen utensils measure liquids crack eggs
You name it they can do it
MasterChef winner Tim Anderson was hired on as Molly’s development chef and
Helped the robotics team teach the robots how to cook by using motion capture gloves and wrist bands much like a videogame
this means you can potentially record the movements of other famous chefs and have a meal prepared for you indirectly by let’s say
chef Gordon Ramsay
moley robotics recently launched a crowdfunding campaign
Looking for investors and potential testers of the prototype kitchen to help them improve the robots ability
Moli believes that they can do for cooking with smartphone devices have done for
Communications and what vacuum cleaners have done for cleaning according to the product video?
You can read more about the moley robotics kitchen at cnet.com
So guys this was about male robotic kitchen, thank you all this was all about this session
so stay tuned for my next video where Alba’s stopped and technical meth which most of you have in your mind like
charging your gadget overnight will destroy its battery life or
Using a phone on airplane will crash it fine
Well, if you enjoy this video, please subscribe the channel and share the video among your friends
Thank you

One Reply to “HOW DO THEY DO IT | Top 10 Applications of Machine Learning | Intellipaat”

  1. Got a question on any of the applications of Machine Learning? Do write it in the comment and you will get a response immediately by our expert. For Machine Learning training & certification, call us at US: 1800-216-8930 (Toll Free) or India: +917022374614. You can also write us at [email protected]

Leave a Reply

Your email address will not be published. Required fields are marked *