Learn Core Machine Learning for FREE | Ultimate Course for Beginners – Transcript

00:00what if I told you that there's a core machine learning course that does not have only detailed explanation of regression analysis but also have b-day lecture notes as assignments with personalized feedback on that and a scheduled learning plan for you well guess what I'm offering all of this for absolutely free hi I'm ayush and I'm currently working as a data scientist at Epic and as demonstrated working experience as a envelopes engineer at

00:26xenomen and a data scientist at Arif and a founder of and termites and I welcome you to our cool comprehensive machine learning course out here the course is designed to give you a solid foundation of machine learning and tradition analysis which is often overlooked in most of the courses available right now we beg it by laying the down work for machine learning so even if a beginner you are able to understand it you talk about what is learning we talk about generalization we talk about inductive bias and a lot of things initially and

00:53then we gradually move on to regression analysis and recover as much as we go in depth and to ensure that we are quite equipped with the Practical applications as well I'm going to introduce two projects by the end of the course which you have never seen like that project before as well because I have worked personally at several companies and I know how we deliver data science projects to them and the same way I'm gonna do it in this course as well but that's not all if you want to maximize your learning opportunity from this course then we are offering free

01:21upgrades to the course regular assignments with personalized feedback to you and a tailored learning plan for you to complete this course in just one month you can simply go on a GitHub page for the course and then get access to the LMS directly from there and all of this for absolutely free one of the major issues which are on a highlight you might be asking here use just for a question most of the instructor takes up only one audio ticking freaking so much of us to teach I tell you that people

01:47forget the foundation people forget the basics of it and it just just strictly goes with Advanced topics like deep learning and then they are not able to secure the job and then they complain me that hey I use data science has no jobs plus actually if you think in a way that the talent is less than the actual jobs because they are not going to pay you thousands of dollars for just three lines of code which you ride inside anyone can do it and then crowd is increasing but the talent is decreasing they're forgetting what's the base of

02:14machine learning what's the base of data science before getting the solid foundation which was used to be given in 90s and all so that's what the course focuses on we make sure that we emphasize the core concept of machine learning and regression analysis so that you can even learn further things on your own our course is designed to put emphasis on core concept of machine learning and regression analysis so that you can tackle critical business problem and solve it the way real data

02:41scientists solve so that's enough of it now let's get started with the course you can directly go to the LMS and then start from there by seeing the tailored learning plan or you can get started from here let me know if you have any notes in the comment section and find any errors or erata in this course please feel free to make an issue on the guitar page for the course hey everyone so let get started with our machine learning content and actually we'll get started with um a bit of discussion on learning the

03:07machines learning intelligence and give you some of the concepts which is required eventually to build up your ideas right so let's get started now so I would suggest you to open up your notes if you have got it um so once you have the notes that you can see on the front of the screen so we're going to start off with what is learning right so let's eventually talk

03:34about what is learning and slowly we'll get into the get into the concepts of you know a very very good examples of learning and intelligence and Etc so the first one is that what is learning so I'll suggest you to write your answer in the comment box if you know so according to what exactly uh what exactly you know about in terms of machine learning we have the training data we take some sort of you know um the data don't worry what data looks like and everything we'll cover that so

04:01we have the data right and then we feed to a sort of a black box and then we get a trained algo that is trained on that data for example assume that you have a man right you trained that man with the with that particular skill and then whatever output after six or seven months you will be getting a trained man specialized in that skill right so this is how your model model gets trained your model gets the data gets the

04:26training from the data and then get it it gets trained by some instructor right I'll assume that in this case it's an algorithm and then after after some time we get a trained algorithm right that in terms of a very simple or and very you know mind understand standing behind any algorithm but let's talk about a very nice example very nice example which is called which is known as rats learning

04:52to avoid poisonous baits so what is baits baits are some sort of you know food or Etc so the formal definition of beets is that it is it is some sort of you know um it is a type of a prey that is used to attract when hunting so um sometimes you we use that baits to attract you know poisonous thing yeah it it is a poisonous thing that you use to attract uh food right sorry a mouse is

05:20or any any other things whatever for example fish we also had the beets right so in this way we have the baits so basically the problem statement the problem statement is that the rats are learning to avoid the poisonous base so how do lats rats learn to avoid that poisonous Spades right so rats assume that rat came and then assume that rat game a rat game and that and then they given a no they there was a food

05:46available where this was a date this was a bait right uh so it gone they gone and then they um it was like smelling good and Etc right eight rat ate this food rat ate this and there's an ill effect on that game right there's an ill effect of that game there's a effect of rat game when they eat this food and and then rats will not eat this food because this is this is the learning mechanism

06:14which happened right now where where according to the past experience according to the history or the mistakes in the in in the past this rat there and he's not on that will no more going to repeat this step again this is one sort of a learning mechanism another learning mechanism which you can think of is assume that you you have done some sort of mistake in your past or a very simple example assume that uh

06:39you go for someone for guidance right why why they are capable of giving items because they have gone through the problems they have already gone through the problems so you go to them so that you get an easy path so that you get an easy uh guidance from them you you so to achieve something in your career right so in that way I'm saying that they already have an experience of something now they are capable of you to guide right just like I do I guide so many

07:05people right so we I got so so many of folks right now so I am a capable because I've already gone through those problems now I'm capable because of my past experiences so whole point to take out right now I have learned what mistakes I what mistakes and what things I did that leaded me to success and failure right that I have well you'll learn that so now that's a learning mechanism for example you learn you know you you study some maths

07:32concept right you study quadratic equations or maybe some calculus concept so when you study that when you study that um what eventually happens is that uh you you take a look at that particular particular concept and once you take a look at the particular concept now you understand with the helpful examples and then you do a problem and if you did any sort of mistakes you you change it according to you know um like ask ask for help or see the solution but to when you s but but when you saw similar

08:00questions you will be able you you will be able to answer that why because you have learned from the past mistakes which you did right because you are not going to repeat the mistakes which you did in the past even while solving problems sometimes it happens as well sometimes it does not right so you you look that's on the learning mechanism that's why we say the failure is a part of success because it gives you the learning phases the learning opportunities to actually go grow forward

08:28so there are several type coaches which we go through the first approach is learning by memorization approach so this is something that's very interesting uh where we we say in Indian language it's ratification and in ratification what what we do we memorize things whatever we we we we want to study so for example I've seen students to literally memorize you know signs answers uh SST answers even computer science answers because

08:54they literally memorize code and they just go and write in the sheet of paper in uh and then they they memorize literally like even mathematical questions however it's so obvious that daily not will perform well but but but over here assume that if you memorization every concept each and every line of your answers will you be able to perform good in examination so even if one line you forget you'll mess up with all the answers you'll start integrating first answer to second

09:20answer second answer to first right so you have to apply so over here memorization literally fails in your daily life you can not literally memorize everything and then go and sit in examination and you expect your that you in school okay it it works till class six or seven but it does doesn't work when you actually grow right so you can now can't literally memorize each and everything whether it be answers whether it be mathematical Solutions or literally anything

09:48so I hope that get that gave you a good sense about what exactly I want to talk on is um learning by memorization approach where we memorize we learn some things by memorizing uh that sort of thing right so now so now what so now what what exactly I'm going to do is talk about some of the uh examples which we have listed out here for you to better understand things cool so assume that you have you you have a you want to

10:15build a system so you might have seen in your email you know uh Gmail where you where you have several photos promotions folder social folder Spam folders and it's an updates folder Etc so you have these folders and there's one folder called spam where all the spam goes in right where all the spam goes in so what are what are those spams right that's question that's a question so now now we're going to build a system you want to build a system you want to build a

10:41system that given the past data right first of all we if if you're learning something to identify a spam you need to have the data you need to have the you need to have some some sort of you know content to understand tantrum right so in this case if you want to build a system your build a system that given an email given an email it classifies that email is a spam or not spam okay that given an email it classifies whether

11:08that email is a spam or not spam right so now that's a question right so you you so now what you so now what you will do is that the first step is for anything for learning literally anything you need content for this so we need the data for it right so we are we are going to collect the collect the collect collect collect the data um we want to collect the data which is where where we have the spam emails right so we will collect the data which

11:36is spam emails out here so we can we collected the data which is spam emails out there now now we have the content now we are now we want our machine to learn from this content so there will be two types of guys first who will memorize this so before we will talk with the first guy the first level mem will memorize this will memorize each and every email each and every email so machine will memorize each and every spam email

12:00right and now now you're now that machine is capable now with that machine is even if if that is 100 correct even if that memorized a very hundred percent correctly right so machine will be able to class will will be will will be able to memorize each and everything now if our an email comes em email comes through this machine and now this a machine is told to classify this email

12:26to be as a spam or to be as a North spam right whether that email is a spam or not spam so you you got a new email and you classify it so how machine will classify machine will search this email into its memory whether this email is available into end to its memory or not right whether the words of this email is available to in in my memory or not right but if not

12:52then we say that it is non-spam but because the data is but because the the content which we learned is about spam right so if if this is not available in that content in his memory what he what the model has memorized from the data then he then the model will be able to see that it is non-spam but if that email is available then it will be spam so what's the problem with this approach what's the problem with this approach right that's a good question what exactly the problem with this approach