You can get the optimal sleep that you deserve time slips away and it’s morning already the sun is shining and you are happy because you had a wonderful sleep now it’s time to go back to work Yuri eye sets up everything needed for your morning necessities water is set to the optimal temperature news is played according to your mood which is calculated using the sentiment analysis algorithms you finish your work and the cab is booked as soon as you are ready to go back to work how did that feel wasn’t it fascinating well let me tell you that all of this is becoming a reality ever so slowly but it is moving in the right direction how is all of this happening it is Biggers of artificial intelligence so what is AI you may ask them well AI is to create an artificial being that can mimic the things that our human does it is to create a human which does not live it exists now what has machine learning got to do with any of this let me tell you that it does machine learning is a subset of artificial intelligence that helps us achieve one particular goal to teach one characteristic of us humans is that we can learn if we are able to teach our computers to learn – then we are one step closer in reaching our goal of the AI that can mimic a human.
so what is machine learning formally machine learning is the scientific study of algorithms and statistical models that computer systems used to perform a specific task without the use of explicit instruction relying on patterns and inferences instead it is a subset of artificial intelligence and when our machine learning models feel we mimic the way our brain works using deep learning we create artificial neural networks that can process information the same way that we humans do but deep learning is a topic for some other time today we will be only talking about machine learning so how can we teach a machine we should have some steps that we follow to get the system to learn let us go through some of these steps now and before we move ahead let me tell you that the machines and computers that we are going to teach a call as models so I will be using the term models from this point forward step one is data collection to teach a model you need to have the data and the resources which need to be used right obviously we humans use books and laptops and so much more that we can study from what about models they need data collect as much of the data as you can so that you can teach your model step two is to prepare the data we prepare the data so we can remove all the unnecessary information that we do not want to teach a model it simply makes sense to teach a model what is really needed step three is to choose the model you understand what kind of a problem that you are facing and so you choose the right kind of model moving over to step four step four is to teach the model piece the model of all the important characteristics and what the data means and all that it has to do step 5 is to test the model.
you have to know whether what you have thought your model was understood by it or not and whether it can solve the problem that you wanted to step 6 is to optimize or fine-tune your model obviously not everybody is perfect from the start slowly and by practice everybody becomes better and better at their work step 7 predict make your train model work and let it do what it is best at doing so those are basically the steps that we have to follow whenever we work with machine learning now moving ahead what are the different types of machine learning there are three types we have supervised learning we have unsupervised learning and we have reinforcement learning but what do all these three mean well let’s move ahead and keep understanding it better let’s teach ourselves so that we can tease the models what they need to do we will firstly start with supervised learning so what is supervised learning supervised learning is the machine learning task of learning a function that map’s an input to an output based on the input and output pairs it infers a function from label training data consisting of a set of training examples so how can we simplify this think of a teacher teaching the students the student knows what they learn and it has also been taught by the teacher now think you want you model to differentiate between a penguin and a pigeon how would you do that it’s very simple you collect data you clean it and you feed it to the computer make it learn from the data and then ask it whether it is a penguin or a pigeon.
it will be able to tell you the difference isn’t that cool well there are some famous models such as a linear and the logistic regression random forest naive bias and decision tree well then so what is unsupervised learning unsupervised learning is a type of of self-organized learning that helps find previously unknown patterns in data set without pre-existing labels it is also known as a self-organization and allows modeling probability densities of given inputs let’s simplify that think of a student who has everything to study from but no teacher what does that student do the student has to study by himself that’s the same thing with unsupervised learning think you have a basket of fruits but you do not know the names of the fruits but you do know that there are different fruits in the basket what do you do next it’s simple you program the model to do the same so that it can differentiate between the fruits in the basket you make the model own so that it can differentiate between the fruits for the color the size and the shape of the fruit so if you find out that there are three major fruits in the basket they are Apple orange and strawberries you group them together is the simplest example of unsupervised learning this would seem similar to an interview process a company comes in knows nothing about you but takes a look at certain abilities of yours and groups you into different departments this is a live example of unsupervised learning we have some unsupervised learning algorithms such as the Keens class ring a priori algorithm hierarchical and Association rules so the last type of learning reinforcement learning reinforcement learning is an area of machine learning concerned with how software agents or to take actions in an environment.
so that they can optimize or maximize some notation of the cumulative reward word is this type of learning in simple words you do not have data to teach nor does the student know how to study the student then slowly tries to study it picks up the pencil it opens up a book it writes the pencil facing backwards walks out of the room thinking that it is called studying it does everything that it can think of but if the student does a good job or an action like writing on the book properly he gets a reward that is how the student knows that he has done a good thing else the student is punished it’s simple as that this here is a game as all of you may know super mario but there’s a catch here this game is being played by an algorithm called as mario this algorithm knew nothing about the game it started moving from block to block using the actions created by a network that fits best the goal what is this goal it’s to complete the game the game tried for ours and ours in the beginning it was terrible at the game but slowly as it learned it understood and became better at it so the agent does an action and he gets back a reward and a state so with enough interactions and training itself it goes from being this bad to being this good in machine learning just amazing now that we know the different types of machine learning let’s move over to the applications that can be useful Facebook uses machine learning to learn about the different people’s on the website and in your photos and recommend tagging them directly learning how to predict the weather from the data collected over the years is such an amazing application machine learning models can learn.
which he means are real and which he means our faith using the nave pass algorithm and so they can act for you as a spam protector virtual assistants like Google and city are trained to recognize your voice only for certain keywords like hey Siri they can do your work that you asked them to do Google has also gone as far as making their assistant make calls for you but here’s the catch the voice is just so real it does not feel like it is a robot can Google be the one to pass the Turing test and bring out an AI that shocks the world only time will tell recommendation systems are another application where association rule algorithm is used to find out what users want to buy and recommend them accordingly machine learning has surrounded us literally everywhere so why not use it to your advantage become somebody who makes such models go ahead pursue your career an average machine learning engineer owns about a hundred and forty thousand dollars in a Europe which is just amazing certain roles that you can pursue are a machine learning engineer an AI engineer a data analyst a data engineer and so much more if you are interested you can check out the master’s program that we have provided at Ed Eureka to kick-start your career today that basically wraps up our session for today hope you could understand the basic idea of what machine learning is stay tuned.