I am Kushal from LearnVern. ( 6 seconds pause ; music )
Today we are going to see
Getting Started with Machine Learning.
Meaning, we will try to learn about Machine Learning from its start, that is from the scratch.
So, let's go,
First of all let us start with the definition of Machine Learning.
Now, What is called as Machine Learning?
Machine learning is an artificial intelligence's subset.
Artificial Intelligence- meaning a very broad thing
As we already know
Alexa is a product of artificial intelligence
Or, there are automatic cars or auto pilot cars
And there are automatic washing machine.
All these are products of artificial intelligence.
So, machine learning is a subset of artificial intelligence.
Now, what happens in this?
Now in this we have a machine, that is a computer.
To this computer we provide some data for its learning,
And then the computer learns it and based on this data it also provides some output.
For example, supposingly we human beings observe some new thing like a game and start playing it and learn it within a few moments.
In the same way, if some system or a computer comes across something new and, experiences it and with this new data it learns something new which also works properly.
And in that 'ours' i.e we human beings do not have to give much interventions also.
So, this is what is known as machine learning.
Now, let's understand it even better on the basis of this diagram.
The first basic requirement is Data.
As soon as you enter the market, you see a lot of stuff.
Now, you see a sweets shop, where you can see a laddu, and also a barfi.
Now, tell me how you could identify the ladu? or, how you could identify the barfi?
Because, when you were small, your elders had taught you that, look this sweet is called barfi and this is a ladu and this is peda.
Now, here if your cousin or your brother who is just 1 or 1.5 yrs old and have just started speaking.
If he starts asking you, and obviously he will, that's what happens when you go with a child he keeps asking, Brother, what is this?, Or Sister, what is this?.
And, now he has started collecting new experiences and new data, meaning he is observing new things.
So, these observations that we make or these experiences that we take.
If we write down these same experiences in a digital format or convert it into a digital format.
So, this is what is known as data.
And as we learn it,similarly if we give the same digital data to the machine, it can learn it too, and even understand it.
And after the machine has already learnt it, it will be exactly like us, to identify them, that these red colour ones are ladus, these are gulab jamuns.
Now how did you know it? because you have already seen it.
In the same way, if we give the data to the machine, then it will also in its output be able to tell you what is a rasgulla and what are ladus.
It can also give you predictions, and also can give you recommendations.
Because, today we can see, how youtube gives us video recommendations,
we can see how online website recommends us different things,
And even,we can also see how the weather forecasting department already can tell us, whether tomorrow it is going to rain or not, whether the sun will rise properly or not, or whether it will be cloudy. And how much the temperature will be.
It can tell us all.
These are all possible only because of Machine learning.
Understood up till now ?
Now, so let's go to understand it in more details.
We have three types of Machine learning,
One is supervised learning,
Second is unsupervised learning.
Third is Reinforcement learning.
So these are the three major types of learning
Although there are many subtypes of learning that can come over here.
Like that of semi supervised learning or on the basis of probability also learning is done which is popularly known as probabilistic learning.
If you mix one or two types of learning then that is known as ensemble type of learning.
So, there can be many types of learning.
So all these classifications would be completed and you will learn everything in a very detailed way in our future classes.
So, now moving ahead, let's see
What is the relation between machine learning and artificial intelligence?
Are they the same or are they different from each other?
If they are different then how are they different ?
So let's understand this.
So, as we had started in the beginning
With washing machines or vacuum cleaners or let's take Alexa itself as an example.
They contain artificial intelligence or they are AI products or, they are AI based products whatever we call them.
Today when we see the description of our mobile cameras we very proudly say that it is an AI based product.
So, what is this AI?
AI is a broad concept in which a machine starts behaving smartly.
Now, see why washing machines are smart ?
Because it can remove the stain as any human being can do it.
So it efficiently does a smart activity.
Now, a vacuum cleaner is there which does brooming and mopping in the same way as how humans can efficiently do mopping and brooming with the help of a broom for cleaning.
So, this is also a smart activity.
If a machine does a human-like activity or does a smart activity then we start calling it an AI based machine or an AI product.
Now, when we move ahead in this and talk about ML.
So as far as we talk about AI based products only, in ML, we don't know whether the machine has learning abilities or not?
Meaning, for example if there is a vacuum cleaner and if it comes across a new type of dust so will that machine clean that dust, by
learning about it.
Or if a vacuum comes across a new type of a thing which looks like dust but it is not dust, so can that machine skip cleaning it.
So if it has learning ability then we will say that though it is an AI based product it also has the ability of learning. Therefore it is also a ML product.
So, this is the similarity between AI based products and ML based products.
Clear or not ? okay!
Now, I will give you an example of a Robot.
My Robot wants me to paint.
So I said ok fine let's do it.
But then he says first show me something. I want to see some paintings.
Then I showed him this painting.
In this painting there are some scenes of a village. It's a nice painting.
Now, he cannot learn from a single painting,
We human beings also cannot learn anything in one go until we practice it twice or thrice, then only we can achieve any expertise in it.
So, Now I told him ok see this second painting, this was a different kind of painting.
Then I said ok now see the third painting as well.
So in this way, I showed him three different kinds of painting.
Then I said, now start drawing it.
Now, see what my Robot has drawn.
Now, you can see that he has not drawn those types of similar paintings that I had shown.
But I am happy that at least the Robot has learnt to identify Colours.
At Least he has put different types of colours there and set something.
Though he has not drawn anything logical out of it.
But at least he learned something.
So, the robot is an AI product, but the things he has learned is a part of ML i.e Machine Learning.
I believe here you can very clearly understand that AI is very broad.
And machine learning is a part of it.
Like human beings have brain, memories.
So the memory captures from eyes, hands, senses and stores in the memory.
And now, to process this memory we have a brain.
So, similarly ML collects the data and understands it, and on that basis it gives the machine an ability to learn something.
Ok, let's move forward towards conclusion.
So, Machine Learning along with other techniques jointly together forms to create Artificial Intelligence.
Now, in other techniques there are some IOT techniques that are used, some data basis, some IT systems can be used.
So, here, there will be machine learning involved but to make the application smart enough and provide it with input and output, you will have to use some other techniques as well.
Now, the most important thing is data.
If there is no data there is nothing.
So data will come, in it there will be pattern, the pattern would be identified and only thereafter a machine learning model is created.
So data, then pattern and then machine learning model.
And after that,
Where and who uses these machine learning models?
They are used by data scientists and in whichever application it is used, it's optimization is also looked after by them.
They are the ones who take care of model improvement.
So , in this way upgrading and keeping the models improving is also the work of Data Scientists.
So, now you can see it's a repetitive cycle.
It keeps on happening, first optimization of
model and then its implementation, then optimization and then again implementation.
Also, we will have to check the results in between to understand it's accuracy.
So, in this way Machine Learning and Artificial Intelligence both work together.
Ok, so now moving ahead.
Let us see what are the tools that we have for machine learning.
Here, you can see that we have Scikit-learn (pronounced: sai kit learn) , pytorch (pronounced: pai torch) , tensorflow ((pronounced: ten sar flow) , weka, knime.(pronounced: na i m)
I have just written five of them here, but there are more than 50 and even more very important and very popular libraries and tool sets that are helpful to us in using Machine learning.
We will learn in great detail and depth in our course.
So, for now, we will just maintain the familiarity with their names.
Now, I will give you some examples so as to make you understand in more depth and clarity as to what we are going to do after understanding this machine learning?
So, you all use maps.
Now, Maps are Fantastic examples of Machine learning.
As soon as you open the map, you say I want to reach this XYZ place.
Immediately the map shows you the entire way to it.
And it also shows about the heavy traffic locations in between your way, and it can also tell you the minutes it will take if you will go walking, and the minutes it will take if you use a bike,and the minutes it will take if you go by car.
So, this is something even human beings will take time to develop this level of understanding and to tell that.
So this machine by collecting more and more data has become so expert, that it can give you near to accurate answers.
So, this is a very fantastic example.
Let's, go towards our second example
everyone uses e- mail
If you are a student,or a working professional or even a business owner.
Email is used by everyone, in whatever domain an individual works, everyone uses it.
Now, what is a fantastic thing over here?
Whenever an email arrives we see that on one particular email it is labelled as spam, the other is quoted as social media, the other as promotion and some are even star marked.
So what are these?
These are classifications of email,
Emails come with automatic classifications.
So, machine learning have become so strong that it can easily identify the emails very well,
It can identify whether the emails are from promotion, or whether it is a necessary email and also if it might have come from a fake account.
So, today's email application gives us a very strong and accurate level of classification.
With its spam filters, smart replies and with email classifications itself,
It has made life really easy for both corporate and personal levels.
Now, I will give you one more example and then conclude with examples.
So, next is Social Networking
Everyone is connected today, especially during the lockdown period everyone was so much connected to one another.
By using various social platforms,online or through conferencing.
Now, see there are different online platforms.
We will take only one example although we have so many of them like Facebook, Instagram and Snapchat and many others.
So let's take Facebook here.
If somebody uploads your photo on Facebook, you immediately get a message that, "somebody called XYZ has uploaded your photo", then you can cross check that.
So, What is this?
This is Machine Learning!
Facebook has so well identified the pattern, your face pattern with its full strength, that if somebody else also uploads your photo it matches the pattern and recognises that it's you and immediately sends you a notification about it.
So these are some of the applications of Machine learning.
Which you will learn and then practice on it and also, you will be able to apply in them.
Our, next topic will be
Environment Set Up
In which, you will learn about anaconda and Google colab.
So, till then remain motivated and keep watching.
Thank you very much.
If you have any queries or comments, click the discussion button below the video and post there. This way, you will be able to connect to fellow learners and discuss the course. Also, Our Team will try to solve your query.
good learning but the content titles are jumbled up, like first title of this module is decision tree dichotomiser which is practical part ahead of theory part. Same with the SVM practical 1 title has
Isakki Alias Devi P
yes, i am happy to learning for machine learning in LearnVern.it i s easily understanding for Beginners.
Superb and amazing 😍🤩 enjoyable experience.
Muhammad Nazam Maqbool
Absolutely good course... will suggest it to everyone. has superb content that is covered in a fantastic way.
super course and easily understanding and Good explaned
Ruturaj Nivas Patil
Very well explained in entire course. Great course for everyone as it takes from scratch to advance level.