 So very good evening to all the educators who have joined us from around the world for the day number five in international bootcamp on coding artificial intelligence and robotics for the school educators. We heartily welcome you all and thanks for coming and joining us on time. We will be shortly starting the session in just one more minute. Till the time, let me just call up my colleague Tina. Hello Tina. Hello, how are you? How are you? I'm doing great. Thanks. And I think all the educators are also doing great. We can see a wonderful comments on the chart. A very good evening to Akansha, Binoy, Balbaster, Ravinderji. So good afternoon, all our dear educators. Happy to see you all. On this fifth day, I would like to welcome you all to International Bootcamp program organized by Stempedia and Art Park. We are really excited to see all your projects in our telegram group. Keep sharing the activities so that we get motivated. A humble request from our side, all the educators, please fill the survey form at the earliest so that we can help you in all your ways. Now let's start our learning journey. Over to you, Ayush. Thank you, Tina. Again, a very good evening to all the educators, Vasya, Krithika, Isitronik, Srinu, every educator. And it's amazing to see that you have all joined on time as well as we are very, very happy actually to tell you that it was really amazing going through with all the projects which you are submitting. So still if some of the educators are not a part of the telegram community, it's a kind request to each and everyone that please join our telegram group because we are using this medium to communicate you to keep you updated with the things which are currently going in the bootcamp. So please scan this QR code. Also, you can click on the link given on the chat box and you can just click on it. It's really very important that you install telegram first and then you click on the link to be part of the telegram group which has been created for all of you. So of course today we are going to learn about machine learning. We are going to discuss all and all and how exactly the machine learning has changed the, I should say, changed the way the computer thinks. Yeah, so moving ahead with the collaborators. So I would request Tina to please let us know about the collaborators, how exactly they have supported us to make this bootcamp successful. Sure, Ayush. International Bootcamp program is hosted by Stempedia and Art Park, AI and Robotics Technology Park by IISC Bangalore. We have been supported by Neethi Ayo that is Atal Innovation Mission, which is a part of Central Government's education project in India. It's been a great pleasure to have all India Educator Forum and American Indian Foundation to be as our community partner and helping us to reach our maximum. We have been supported by many international partners like Purpose Smart Education Palestine, National Institute of Education Singapore, Apex Coding Academy, Egypt, ITA, Tunisia and Edustream UAE. All of our collaborators Stempedia and Art Park have made a wonderful effort to bring all of our educators. Now, it's time to discuss about the objective behind this bootcamp program. The objective of International Bootcamp program is to empower the educators by deepening their knowledge on coding, artificial intelligence, robotics and thus enhancing them by giving all the necessary skills that they required on this 21st century. By this bootcamp program, all the educators are getting a hands-on experience on the new day technology like robotics, artificial intelligence, machine learning in a fun and interactive manner. By learning more about the new day technology like coding, artificial intelligence, robotics and even machine learning, the educators will be confident enough to talk about this technology in front of our society. This bootcamp program will definitely enable the teacher to become an ambassador of their own institution. And now it's time to see some wonderful creation of our educators. Ayush, you can take it forward. Thank you. Yeah. Thanks, Tina. Of course, before showcasing about the projects, I would like to tell you some important points educators. Of course, you have joined here and you can see in the objective, there is a last point that is enabling the teachers to be the ambassadors of robotics, AI learning in their respective educational institutions and in the society at large. And this is one of the major objectives, what we want to have. So we would be talking about that how exactly together all you educators, our team, as well as some of the special guests, which would be coming in the closing ceremony are going to help you to do so. We would be talking about during the schedule. So let's move ahead with a wonderful creation. So these are some of the wonderful creation. And it's amazing to see the creativity of all the educators who have shared their things related to the virtual doctor, as well as the NLP. So I hope you also are going through the lessons you are submitting your assignment. It's really necessary that you submit your assignment, you submit the videos you share the things in telegram group as well. Now moving ahead with the things which we wanted to take that why exactly we all are here and we are learning these things. Of course, one of a major point here is that we wanted to learn the 21st century skills. Also, you I think maximum of you have already joined the telegram group. Of course, that's an international educator community. As you are submitting your assignment that is making you qualified for earning a badge in the certificate accredited by art park stampede and stem.org. Still, if you have not submitted, you will be having a time for tomorrow that you can take your time. And of course, you can submit the projects I should say, just by making the video of the activities and sharing the YouTube link in the activity part. Apart from it, as we are going ahead in the bootcamp, and we as we love to always fulfill all our promises. So there are two important things which you will notice in the screen right now is get up AI and robotics teaching resources and access the students learning resources. And of course, as I mentioned that we want you to be an ambassador of the AI club in your institutions and the society. And these three combining these three, we are going to make the complete session, which is going to coming on the Sunday that is on closing ceremony. So we highly recommend and we, I would say that we would request each and every educator to be part of the closing ceremony, because it's not only we are just going to talk about all those things, but we are also going to let you know about the projects and the batches how you can acquire them after completing the courses. All the resources would be shared in the coming session. So it's really necessary that you complete your activities and be a part of that, be a part of the closing ceremony, so that the most important objective as an educator is not only to learn, but of course empower the young So with our team, we are going to help you to know the complete execution plan and surrounding you the different type of resources which would be required by you in becoming the ambassadors of AI in your institutions, as well as in your society. So now moving ahead with the bootcamp details, so I would request Tina to please take over and let us know about the overview. Sure, so as we all know, this bootcamp program have two levels that is level one that is coding and artificial intelligence in that we covered the first session introduction to artificial intelligence we learned about a phase detection, we did the activity of emotion detection and all in session to object detection we learned about object detection techniques, and we did an object tracking activity. In session three, we learned about human body detection we learned about human forces different human And we made a game which is controlled by our hand. In the fourth session we learned about natural language processing we discussed about text classifier, and we did an activity of text classifier. And in today's session that is session five will learn about machine learning will understand how exactly machine learning something and will do an activity of animal classifier. And in the sixth session that is on Sunday, keep in mind like tomorrow we don't have any session in session six, that is going to happen on Sunday that is on 26 June, that is our closing ceremony. So as I said, this closing ceremony is going to be a very, very important part in the educators journey in this bootcamp. As till now we were always coming to the session and we were learning the things. So how it's our responsibility as well as it's our main motive that how exactly the you can implement and make these things transfer to the maximum students and become an ambassador of a so this is very important as I said that you have to be part of the closing ceremony, and we are also going to discuss this discuss about the batches and the certificate how exactly you can share them how you exactly you are going to earn them. And that is all I can say that you have a one more day to submit your assignment and on 25th of course you can take your time. Take the videos wonderful videos submitted on the activities join us on 26, make yourself ready to just rock the world with the knowledge what you have acquired here and transfer it to the young mind. Now moving ahead with a schedule of the bootcamp level two. Number two, we have also we also have six sessions and level two is robotics and artificial intelligence that is going to start on 27th of June. The first session is introduction to robotics will learn about some basics of robotics physical computing and will make a robot to move. The second session will learn about self driving robot will learn how sign detection working with the use of artificial intelligence and will understand the concept behind the self driving cars. In the third session will learn about line follower robot will learn how to calibrate an IR sensor will understand its logic and will make our robot to follow a particular path. In the fourth session will make our robot as a delivery robot will learn about some basic applications and its logic. In the fifth session will learn about guest control robot will learn how poses are made how poses identified by machines and made by using machine learning and will make our robot, which is controlled by our guest. And finally in session six is our closing ceremony and that was a complete schedule. Ayush, you can take it forward now. Thank you. Yeah, thank you Tina for like giving us an overview of the schedule. So I think you can just help educator by sharing the telegram group link on the chat. So I think some of the educators are not able to join not a problem we'll just share you the link. Stay tuned with us till the end of the session. Now of course till now, whatever we have discussed whatever we have covered still if you have somewhere you have not started or you have missed the session, you don't have to worry because right now as you know these sessions are going live on the YouTube playlist of the Stempedia, you can always come back and see those sessions, but we would highly recommend you to watch the sessions in your own dashboard at ai.thestempedia.com where you can just log in by signing up to the ai.stempedia.com and enrolling into the boot camp. Today's session you can see this is a once you have reached to the international boot camp part, you have to click on my courses. Once you have clicked on my courses, you would be able to basically see all the enrolled boot camp. So here this is the boot camp which we are currently working on. So this is AI coding boot camp for educators level one. Now here, as I said, we are still going on with the doubt session where you can join us and of course after all questions from 6 to 7pm IST and 10pm to 11pm IST. So please you can join after you can join the session after the live session and you have to go through with all the content here so that you can meet the activities as well as you can submit them. So today, of course, we have a session of machine learning. Okay, and here you can see all the forms are available. You can just fill them attendance forms. So these are the attendance forms which you have to fill. We would be talking about these things little later as well. Now, coming to the part, the activity. Still, if you have not submitted the activities in every session, then you would be able to see as activity one and activity two. Okay, so here in activity two, once you'll go here, you would be able to see a place to submit your YouTube video link. And basically, you have to name the assignment like today we are learning about machine learning. I'll name the assignment with Ayush machine learning. Okay, and I have to paste the YouTube video link of the assignment. What should exactly the YouTube video contain should contain is your YouTube video should have a little explanation of the concept what you have learned in the live classes. Your video should also contain showcasing a coding part as well as showcasing the final working of the project, because this is not going to only help you to make the help. It is not only going to help you to make the projects and recorded but it is also going to help you are your students when they would be seeing your videos on the YouTube and they would be very cherish to see their own teachers on the YouTube. So I hope you all are submitting your assignment. As if you are not able to join the telegram group, let me just share my screen back again. And let me just show you the QR code. So this is the QR code which you can of course scan and of course I should say join the telegram group. I hope we are all excited for the today's session. Yeah, sure. If you have uploaded the video link without the name not a problem although it is going to be with us with your username and email ID. So that is not a problem. Of course, all of those things are accepted, but we are trying to just follow a convention of naming the file so that of course, even you would remember we would remember the thing. That's great actually need to do for submitting the daily YouTube video link, and we have been watching the videos by all the educators, and that has always motivated us to bring the new thing for all of you. And of course it is going to be an amazing journey towards where we are reaching. Okay, so now we can start with our today's session that is about machine learning. So I hope we are all ready with with your with your devices in the devices with pictoblocks and of course excited to learn the machine learning. So let's start with an introduction to machine learning. The first and a very most important thing to know is how human learns, and then we would be able to relate it to the machine learning. As we did earlier that we knew how we are saying that the humans are intelligent, intelligent, and then we were learning how exactly we made an artificial intelligence. We made our machines artificially intelligent. Now talking about machine learning. We are going to learn it in a similar way as I said, so let's discuss this small, I should say learning pattern of a human, and I hope you would all be agreeing to this, because I would be taking an example from our own childhood. Okay, so here you can see, basically there are four different steps or stages of human learning process. First is, since the environment, second is analyze the information, third is decide and act, and then comes up increasing the knowledge. Now let's take an example from our own childhood, and I hope I think that everyone here might have faced this. So if I talk about how exactly the learning journey of a child starts. So when the students are in very early age, okay, so they start learning from the picture book. So what exactly are the picture books? So picture books generally contains only the pictures and the name of particular object or a thing. And the learner or you can say the small student or a child sees, generally see that pictures and start understanding what it can be. So generally parents and teachers tell them see this is something looking like a cat. These are the different animals. These are the different vegetables. These are generally the common pictures in the picture book. There of course, the children don't know about what is the animal as well, but still they are able to save that picture in their mind. Okay, and that is you call as sensing the environment. So they start understanding the things with the environment they are able to see in the picture. Now what happens is the children grow up, they are taken outside. And if you remember you would also say yes to me that our parents are the first teachers who let us know about the things. Okay, the cat or the dog which you saw in the images, you can right now see it in real. And they let you know this is the same dog. Can you see, can you make a similarity between the dog which you saw in the picture book as well as right now you are seeing it in real. And there what happens the brain of the children starts working. And the first thing what it does is start taking out the similarities. Like there is a high chance that the color of the dog would my I should say would not be exactly the same in the picture book. But what happens the different features would remain the same like the ears, the nose, the body structure. Okay, so now children start taking that lesson. Okay, and you can say it like analyzing the information. Apart from it, what exactly we understand when we see the different animals in life is we start analyzing their behavior to understand this example. Let's take two animals. One is cat and one is dog. So if you might have seen cat or dog. Okay, so what happens is you might have observed their different behavior. Like you can say dogs are sometimes little more aggressive. Cats are not cats can climb the walls. They can climb the trees. But of course dogs cannot dogs are somewhere if you feel they come and chase you while cats run away from you. So these are the certain, you can say are different features or different behavior of the animals, which are children or a new learner observe and analyze it. Now after analyzing certain features, the part comes off making a decision. Okay. Now, how exactly the decision is taken. So if you'll, I'll just take you in a situation. So it has happened actually with me and I think in when we were very young it happened with almost everyone that we were going on roads and suddenly a group of dog comes up. Okay. So there we have to take a decision that should I go ahead, should I run from here, or what should I do. So if I ask you, and if I give you this scenario, what you are going to do that if you are going on a road, and suddenly a group of dogs, okay, and came up and stops you or you can say hinder your way. So what you are going to do. So you have to particularly take a decision based on the knowledge we have acquired till now. I would love to see some of the answers on the chat. I'll repeat my question that if you're going somewhere and suddenly the group of dogs comes up and so what you are going to do at that moment. Some of you would have first going to say that will run will just go walk slowly. Okay, but if I say that my decision would be going back to my home. Okay, I would not go with that particular way I'll come back to home I'll think. Okay, this was the way where I was wanted to go but I just got hindered by some animals. So now what I have to do is I have to take a second way. Right, and that is a decision which I took, and that was all completely based on my knowledge of the dog behavior that they can chase me they can of course, sometimes also bite. Okay. So now what happens is, let's assume I took a second way and I went to my friends home. And then what happened suddenly for the first time I saw a dog. Coming and after coming towards me, of course it was a pet dog. But if it's the first time for me to see that pet dog, of course I'm going to get scared of that too, because I don't know about the pet dog. At that moment what happened, my friend told me you can beat calm because this is a pet dog and it's not going to harm you can spend time you can make love you can just feed. And you can play with the dog. And they are exactly what happened. There was an increment in my knowledge that now I just got to know that all dogs are not seen. There are particularly different breeds to whom we can make them pet and if we can of course I should say have them and we can take them in our homes as well. So now this is something what I did was an increase of knowledge. Now, no, this slide is not changing actually. So yeah, so we are on the fourth stage that we are increasing of where I increased my knowledge. Okay, now I'll take you to a part where I would like to tell you that the third and the fourth stage is always going to work, because there might be a chance that you do not do didn't took the right decision. And still there is not a problem. If even you have not taken a right decision that is going to be a knowledge for you. Okay, that is going to be a knowledge for you that next time. This has not to be done. And this is how the third and the fourth stage goes on and goes on. Okay. Now, coming to the next part that how exactly we also actually I would take you back to the actually I was not changing the slide. These are the only four stages which we were talking about. Okay. So now coming to the part of how exactly we increase our knowledge we make ourselves comfortable with the knowledge which is coming to us. So we have of course different inputs you can say like vision, touch, smell, taste and sound. And here also, let's take some example. If you'll think that like if you currently if you are currently sitting in your drawing rooms. Okay. And suddenly a voice of barking a dog comes from outside. So there you are not going to even see what animal it is. Just by sound, you would be able to detect. Yes, this is a dog. And that is what we are calling it is that we have learned so many things about a category of animal that is a dog. Okay. Similarly, if I say you want to buy some sweets and you go to sweet shop. So does this happen that you taste all the sweets all over the switch shop. Of course not just by seeing the sweets or seeing the cake you are able to feel okay what could be the taste of that particular dish or you can say that particular sweet. And you are able to directly audit. So that is what we are calling it is you are able to grab a knowledge you are able to track it in your mind and you are able to take the decision based on it. Now, how exactly this process work was we analyze the information we decided and act upon it and whether the decision was correct or I should say incorrect. Still, we were able to increase the knowledge. Right. Now, if we talk about machine learning that how exactly machine learn the machine learning of course have a two word that is machine and the learning. So, combining combining them both we can say we are going to talk about how exactly the machine learn. So machine also learns in a similar fashion that machine learning process is the process where exactly we give the data we ask them to understand it and then we take the output right but let's understand it with a diagram. Okay, there are three basic steps. First is input. So in human case input was of course the picture books, the knowledge by the teachers and the parents, but in case of machine, the input can be very limited. That can be an input of an image sound, okay, or pose. Now coming to the part once you have given an input as a, I should say once you have given an input as the images or the sound, then the second part comes up of a model. So basically what exactly the model is the model is basically an algorithm which is operating in the backside of the complete machine learning process to make your machine learn this concept. Okay, and of course so that these input which you have given should be understood by the machine to get the output. And that is very important of course we are all doing the things to get an output from that right. Now we are talking about output is the machine should be able to make the decision automatically. Once if I have taught how exactly the cat looks or a dog look my machine should now be able to just see any of the images of cat or any of the images of the dog in with any of the color variant, but should be able to tell me yeah this is that. So this is how exactly we want our machines to learn. Now how exactly we are going to do it is we are going to use a platform of teachable machine. And let me just tell you that what different types of models we can make. So basically we can make an image model to talking about what exactly is image model that means that we are giving input as an image right if you are giving an input as an image that makes it image model where machine is learning only through the images. The next is of course the post where basically we'll see I'm sorry. So where basically we'll see that different poses are the machine is learning different poses, right. So we have already know we already know about this thing we have understood in human body detection that the poses comes out of with an 17 different points, but of course we have to teach them so there were certain. I should say, there were certain requests from the educators, they wanted to know that how exactly we can train the model. And today we are going to do that. So I just will start with today's activity. So as today's activity is that we are going to make a project on cat versus dog that my machine should be able to detect what is that and what is dog. So let me just switch on my picture blocks. And I hope till here if you have any doubt related to concept, please put it forward on the chat I would be able to answer them all. And if it's completely clear you can, of course, yeah you can of course let me know that the concept we have talked about is completely clear. So I hope you would be all able to see my picture blocks screen now. Now if it is visible please let me know on the chat that it's completely visible. Okay, thanks thanks teachers. So just give me a sec I would also request till the time you can just switch on your own picture blocks. So here we would be working on making a classifier for cat and dog. So I think like, still as a human of course we are very able to classify the different animals. But for a machine it's always a new concept if you want them to make to like make a difference between the animals, right. So what we are going to do is offer the person a very important part here is we want that an extension which can help me to do so. Right. So can you please tell me if I click on add extension option, which extension I can choose to go ahead and make to go ahead and make the project. Okay, so yeah of course we are going to take a machine learning extension. For this extension, it is really important that you have internet connection, as you can see that when you are working here you requires an internet. So if if picture blocks is taking a little bit of time. You would require suggest you to like do you stall it because sometimes what happens is what the file which you have installed a picture blocks might have not got stored correctly. Okay, so now what we are going to do is the first thing I hope my screen is completely visible and now what we are going to do here is of course we need to first important thing is create a model. Okay, so once you have to once you have chosen this option of I should say machine learning extension here you would be able to see there are no blocks added in your screen. So you can just think what why exactly there is no blocks right now on my screen, even I have added an extension. I tell you why exactly it's not here, because still the time, whatever extensions we were using was already there inside the picture block. Okay, so the models related to them were already the part of the picture block. So when you added it, you got the all the blocks in here on this stage. But right now, this is an extension as I said one of the best extension as well as one of the very powerful where you would be able to customize the things as per your own understanding. So what we have to do is you have to click on create a model. Okay, just click on create a model. So once you have clicked on create a model. So basically you would be directed to this teachable machine platform. So this teachable machine platform here basically gives you an option to create three different types of projects, although we are not going to work with audio project. Okay, because audio project is not such right now, you can say is working with pictoblock, but still we are going to have it soon inside the picture block. So now what I have to do is I have to click on image project. Okay, so once I click on image project, since we are going to make a classifier. Right, if I think of a classifier, that means I have to first give a data. Okay, so let me just click on image project. And here we are going to use a standard image model. Okay, so here basically what we have to do is we have to click on create standard model. And once you have clicked it here, then the next part is you have to give input data. Okay, now here, before we go ahead and give an input data, let me just give you an idea what exactly the setup is. Okay, so if you'll see it here, as we understood that there are different steps while in you are doing the things in machine learning, that is first, you have to give the data, or I should say you have to provide the data for machine to understand. Then the second part is you have to ask the machine to get yourself trained. Then you have to test it. And then the next is you have to deploy that. Right. So how you can do it here is the first is input. So the first two things you can see class one and class two are basically a place where we are going to have the input. Now what are classes. So when we're talking about classes, it's very important we have to understand it with like, it's a label. Okay, so we have also worked on classes just today when we were working on NLP. Right. So here also we have to give a name to the class. So right now what I'm going to do I'll just write the name of the class as dog. Since I would be making a project or I should say making a classifier on cat and dog. Okay, now the things come up. Can I do this only for two classes or only for two enemies? Of course not. You can add it as many you want, but we would highly recommend you to keep it only the around seven to nine because it is going to somewhere in the resources from your computer and so make sure that you are like using it very properly. It doesn't enhance your PC. Okay, so you can have at least a very minimum I can say five to seven classes you can use very easily there won't be any problem. Okay, so as of now I'll just delete it. Now of course there are two options here. One is upload and one is web camera. If you're going to click on web camera, let me just show you if you're going to click on web camera, it is going to switch on your web camera to give an input of the data. Right, but since I want to work with cat and dog classifier, so I cannot give it directly from the camera as I don't have both of them with me currently. So what can be done on that part is I can give it by uploading that data. Okay, so how can I upload it here is by clicking on upload, but I should also have as I said, I should also have the data set. Okay, so when you would be doing a course, let me just show it to you. So here basically you would be getting an option to download the data set. Right now we are giving you the link on the chat from which you can download the data. So this is an option. Once you click on that link, the zip would be downloaded. Okay, a zip file would be downloaded. I'll show you. So that zip file would be somewhere getting downloaded in your downloads. Okay, and then you have to just open it. So you can see in my downloads, you would be able to see the zip file. Okay, I'll just open it. I'll use some of any of the unzipped, I should say software. I'll just click it here, expect here. Okay, so just a sec. Okay, so now I have downloaded and I have also unzipped that file. So I have to go back to my teachable machine. Okay, and now from here basically I have to upload it. So I'll just now click on upload. Okay, and here again you have two options. One is you can upload it from directly your machine. Okay, please share the zip file. Yeah, I'll request Tina to please paste the link for the zip file. Okay, so once you can even choose it from your local drive and even you can choose it from your Google Drive. So yeah, we would be just shortly sharing it. So once I'll go I have saved that folder already in my desktop. So I'll just go. Okay, and now here you are going to see two folders inside this dog versus get you are going to see two folders. One is training data set and testing data set. That means basically we would be right now training my machine. So I require training data set. I'll tell you what is the difference between them. So training data set it has two folders again, that is dog and cat, because we have kept the images of dog and cat separately. So I will click on dog. I'll select all the images and I'll open it. So all the samples of the images would be getting uploaded here. Now similarly, I'll just click on upload here. Choose a file. Okay, and this time I have to upload the cat images. So I'll just go in cat and I'll open it and I'll load the sample. Okay, so just let me check with this that just a sec. Let me just share you the link of data set. Yeah, I'm just sharing it in a minute. Just a sec. Okay, so I hope you would be getting this link shortly in your chat box. Yeah, so coming back to the part that coming back to the part if I have I should say uploaded the data that is uploaded the images. And I have also labeled the cat and the dog. Now the things come up offered training the model and that is. Yeah, that is something that is what we call is clicking on training model. So once you click on training, the model is start getting trained. Okay. So now you have to understand what exactly is training right now. Okay, so you don't have to worry you do not have to match my face. You don't have to request because of course this video will be with you always. Okay, and let me just switch off the input first and let me just keep it on here. Yeah, so now you don't have to match my I should say pace as I said this video would be with you and still if I try my level best to keep going with moving ahead with all of you. Okay. So now once I've clicked on training what happened. So there was a model training going behind it. So now it's very important that you have to understand that what exactly is training. So what happens here is the machine take out the similarities between the images in their own class. That means it is going to take out the similarities between all the images that is in the dog class, all the similarities between all the images in the cat class. The next it is going to do is it is going to find out the differences between cat and dog. Okay, the images of cat and dog. Also, if you might have observed the training went from one to 50. That means I have basically trained my model 50 times. Okay, that we are calling it as epoch. So you would be able to see a small number coming up that would be moving one out of 52 out of 53 out of 50 that we are calling it as epoch. That is, we are making my machine basically learn this concept 50 times. So I think everything related to the train training the model is clear. If you have still any doubt, please put it forward in the chart. I would love to answer them all. Okay, now coming next to the part is I have to test it whether my machine has learned correctly or not. So what I'm going to do here is I'm going to choose an image. Now this time I have to use the images from the testing data set. So here you will see we have a mixed pictures of cat and dog and a very important part I would like to tell these images are not were not the part of the training. That means if you see in these images, these that there is no image certainly same like this. Okay, there is no so you can see there is no white cat, but still my machine is able to detect the white cat as a cat. Even though if you see there is no white dog, but how exactly it is able to do so. So that is how the machine learning algorithms work on the basic structure of the picture what you have given for the training. Now similarly you can test this again, you can just choose the any other image from the dog and let me just choose a white dog. So here you can see it is able to detect the dog as well very perfectly that means my machine is completely trained. Now, also I would like to tell you that if you would be able to see the output as cat and dog, which is showing 100%. So what exactly this person means. So that means a confidence level of the machine to give this answer that machine is 100% confident that this is a dog. There might be certain chances when your machine would not be so confident to tell you the answer still it is going to give you the answer. Right, but in that scenario, you have to give little more data of course to make your machines work better. Also I would like to tell you the number of images right now given while training this classifier was 10. But if in case you would be working with some real time projects like mask detector or making different classifiers, you have to give little more number of images. The number of images, which is going to work in the teachable machine best would be nearly about 150 to 200. Okay, so I would suggest if you are making your own project like mask detector, you have to give the number of images around 150 to 200. Now does the model is ready, but of course we want to make a program right it's like we have trained the model. So now to make the program and take this all information back to the pictoblogs I need to click on the export model. As soon as I click on the export model, a new pop up comes up, which says me to upload my model. So upload my model. So basically upload my model refers to uploading all the calculations to the teachable machines drive. Okay, you can say and it is going to give us a drivelink. So right now if you see there is no drivelink here. Okay, make sure that you are able to see there is no drivelink. So here if you see it is dot dot dot. Okay, let me just mark it if it is possible for me. You can see this is dot dot dot. So this is not a link. So now I will click on upload my model. Once I click on upload my model, there would be a dialog coming as uploading. And once the uploading is completed, I would be able to receive a link. So now you can see I've got a link here which says your cloud model is up to date. Now you can use this link. So I have to copy this link. Okay, I have to now take this link to the pictoblogs. Okay, I'll just open up. I'll go to pictoblogs till now. What we did was we created the model and this time I'm going to click on load model. Once I will click on load model, a new popup would come up here where I have to paste it. Again, you have to see that right now there is no link pasted. Okay, so here basically you have to paste the teachable machine link. Okay, so you can see if you are not able to click on the link that is there is no link. So you have to just paste that link. And once you'll be pasting it, you would be able to move your cursor over the link. Okay, if you're able to move a cursor over the link, that means you have pasted the link. And now you'll click on load a model. Once you'll click on load a model, then only your block would be coming up here. Now I would like to tell if you're doing this on smartphone, you won't be able to create your own model, but you don't have to worry, but you would be able to load the model. So I'm just sharing you this model link for those who are working on smartphone. Okay, so this is the model link which I have shared. Okay, so I'll just share this. So this link is for those who are working on a smartphone, you won't be able to create the model. So please you can use the model link which we are going to paste it shortly on the chat box. Now the next is, okay, so we'll share it to you. Okay, we'll share it to you later and even we'll share it to you on telegram group. Right now we have disabled the chart. Okay, so now what we have to do here is we have got all the block, right? So now what the next is we have to make a script to work with the model what we have trained. Okay, so now what we are going to do here is I have some blocks. So let me just show it to you. This block says, let me just zoom in it says is identified class from the web camera. So I'll make it a stage. Okay. So you can see that we have got the same classes here on which we have trained our model, right? On which we have trained our model. So now what you have to do here is you have to take of course two blocks similar to this. We would be checking it on a stage. Once we would be checking cat and once we would be checking the dog. But now the problem comes where exactly from where I'm going to bring those images of cat and dog. So for that I'm going to upload the backdrop and this is you will click on backdrop option. You are going to click on upload backdrop. And here I'll go to my desktop. I'll go to cat versus dog. And here I have to upload all the testing data. Make sure you are uploading the testing data because of course if as I said if we have trained my machine on any certain pictures of course it would be very easy for it to answer. But of course we don't want that. We want it should be able to answer me new things. Then only we can say my machine is smart, right? So now I'll just go back to click on Toby and let me just zoom the screen. So now with this of course we would be able to get the answer. But we have to make a script as well. So let's make a script by bringing the first important block that is when green flag is clicked. Now when green flag is clicked what exactly I want is it should start checking the thing. So I would bring a forever block here. And in forever I would require to create two different conditions. So I'll just bring this block. If the identified class from the stage is dog what I'd want to do? I want my Toby to say this is dog. And I'll put it inside the forever. Now similarly I'll make one more condition. And I'll make it if this is a cat what I want exactly to do. It should say this is a cat. And let me just copy the same part here and make it cat. So this is a cat. Now if I put it here and if I press it. So let's check what it does it say. So you can see it is saying this is a dog. This is a dog. But I wanted to make this little more creative. So that the backdrops which I have added should get changed automatically. So what I can do here is I can go to look. I can just bring a block with the next backdrop. I can add a weight here. Okay. So with every backdrop it is going to check. I'll add a weight here at the end. Okay. Now I'll just press this green flag. This is a cat. This is a dog. This is again a cat. This is again a dog. One thing which I wanted to do is I'll just go to stage. I'll go to backdrop. I would remove this backdrop that is a plane. Okay. I don't want this. I'll delete it. I'll go back to Toby and block. And let me just zoom it. And let me just click it here. All clear with this. But to still give you a small recap, let me just tell you what we did till now. We added a machine learning extension. Okay. And then at the point when we added a machine learning extension, we clicked on create a model. Okay. The first step what we did, we clicked on create a model. Now after clicking on create a model, I was somewhere into teachable machine. So let me just show you. So this is a teachable machine platform. Let me just click if you can get started. Okay. So I was directed to this page. Right. Once I was directed to this page, I had an option to make post project or image project. So what I did, I clicked on image project. Now I had, I have chosen this standard image model because, because if you'll see, if you'll like read whatever is written here. That says embedded image model is for microcontrollers that of course I don't require. I'll use standard image model. Now this was the particular view what I got. And we have understood that these classes are nothing but the label, which we can of course name them. Now second part is we have to give the data to them, right, which we can give them by using a web camera or I can give them using an upload option. We already had a set of training data. Okay. And the testing data with the link which we had provided. Now here basically what I did, we uploaded this image. We uploaded the images of cat and dog. Once we uploaded the images of cat and dog, we clicked on training the model. Okay. What exactly was the training model? So behind this algorithms were taking the similarity and differences between the images of dogs in the cat and came up with a certain solution that what could be a cat and what could be a dog. Now here the important points come up that there were two different folders. One was training the data and the second was the testing the data. While training, we have to make sure that we are giving an appropriate images. That means in place of if we are putting up the images of dog, we should only give the images of dog. Even by mistake, we cannot give the images of cat there. Similarly, we are uploading the images of cat. We have to make sure that we are only uploading the images of cat. Now, once we have trained our model, we tested the model by giving up the images in teachable machine in testing part. And I was able to get the testing results at the bottom of the output that yes, this is correct. This is a dog or this is a cat. Once I was confident enough with this part, yes, my model is trained perfectly. I click on the export model option. Now export model option give me an option to take the link of all the calculation to the pictoblog to make the script. Now, right here, we updated we uploaded the model. We copied down the script and then we came up over here on the pictoblog. As I said, as a teacher, as a if you as if you're using it on your Android phones or in your smartphones, you cannot create basically the models. But of course, you can use the model link, which is which I should say would be shared by us shortly to you. You can directly load up the model by the link. So there you have to click on load a model. Once you click on load a model, make sure you have pasted the link and click on load a model. Once you click on load a model, it will take a time of 15 to 20 minutes. I should say 15 to 20 seconds and will come up with all the blocks which you require. Once these I should say required blocks have came up. Once the required links are came up on the I should say pictoblog. Then of course you can make a script and then you can of course move ahead by uploading the different backdrops of chat and doc from the testing data to check whether the things are working good at your point or not. Still it is not working good. You don't have to worry. You only have to basically I should say retrain your model with more number of images. And as I said, when you would be working with certain I should say with certain different classifiers like mass detector and any other, we would highly recommend you to give. Give the number of images approx 150 to 200 to make your make your models working good. Now, once you have discussed about it, let me just take you to the next part of applications of machine learning. Now there are many applications like object and animal classifier. You might be thinking why how exactly it can be useful. So in the different forest, what happens is you make sure that no other animals comes into the different place. So that like I should say I can say in the forest basically not an animal classifier but a human classifier work good that no person comes in the forest and I should say kill the animals right or hunt the animals. So these different animal and the human classifiers are going to work in different forest areas identifying the shape. So even you can train the model with giving an input as a different images of shape. One of the DIY activity is of course the mass detector. So right now machine learning is also getting used in different services purpose whether what whether the person is wearing mask or not. One of the major use cases of machine learning is in healthcare sector. Basically, here you can see even you can make your machines learn about it using the X-ray images that whether a person is affected by pneumonia or not. So there are many other use cases even healthcare health sector where nowadays using the machine learning the I should say the softwares are able to detect pneumonia cancer and many other diseases. Even you can make wonderful games where you can make a rock paper scissor and plate with the machine where you are able to showcase your I should say hand images and you would be able to get the images output from the machine. Now this is a DIY that is do it yourself as I said a machine learning using machine learning you have to make a mass detector. Still if you case you are finding any trouble or any problem please feel free to put it forward on the telegram group we are going to be helping you out there. Okay, and a very important part do not forget to fill up the feedback and that in inform that is as I said. So it's two different form one for all the educators from all around the world in India, while there is a separate teacher community from the Niti IO game, we request them to fill a separate form. It's really important that you share your feedback your share your. I should say attendance as well as we also want you to share that what exactly you are more looking forward to learn and and what type of support you require from our side. As I said, the learning session of course is completed today, but of course you can join the doubt session which is going to be from 6pm to 7pm and 10pm to 11pm. For link of course the link would be shared in the telegram group as well as you can join it from your dashboard. A very important part before I take the doubt is that as I said this is not the end of the bootcamp as we still have one more session which is a very, very important session coming on Sunday where we are going to help you to get your certificate, get your batches, get your all resources so that of course you can use it not only for the students, but of course for your personal and professional development. We are going to come up with a complete plan to help you to become an ambassador in your own institutions of the AI. So now we would be taking some doubt so I would request Tina to please like if we have collected some doubts you can please put it forward. Sure Ayush. Ayush the first question. Can we still enroll to level 2? Yeah. So this is of course a question getting asked by many educators so I would like to tell you a very simple way out here. See enrolling to the bootcamp you can of course go ahead and we motivate you to do an enrolment and get your kit delivered. Since this all learning management and all the videos would be there with you for complete two years right. So you can order your kit you can get your kit there is there might be a chance that kids don't deliver on the first or the second day of the bootcamp. But you don't have to worry with that because we would be there with you till the end of the bootcamp and you can use the videos which are already getting embedded in your dashboard to learn the things and make the assignment. This complete process is going to help you to complete the bootcamp and earn your batch and certificate. So the answer is completely yes you can still book your kit you can still book for the bootcamp level 2. Still if you don't receive the kids on the first or second day don't worry the resources are with you the places with the dashboard is with you when you get a time you learn it you make a video you uploaded there. And you would be earning all the batches in the certificate. Yeah. Okay. We have one more question. Why do wrong credentials pop up displaying after registration in pictoblocks. Yeah. So for the pictoblocks let me just make it very very clear it's very important. The first thing right now you can see this is my email ID. Okay. And this is my username. So you have to make sure that you log in with your username not email ID. Now how can I get your how can you get your username. So the email you might have received while verification have your username on the top. Okay. It says hi. I use under core AI. Also very important part. I'll tell you. If you have registered on the bootcamp on the website that does not make you register on the pictoblocks pictoblocks and the website have a completely different databases. Okay. So if you have to register on pictoblocks separately by clicking on register option. Once you have registered here then only you would be able to use the username which you have created in pictoblocks to log in into the pictoblocks and use all the extension. Yeah Tina. So yeah. Do we have more questions? That's all Ayush. Okay. So that's superb. I think teachers have definitely liked the today's session and they have got a good idea of machine learning. We would be looking forward with you all submitting your assignments in a form of YouTube link in the dashboard. And we want you as we said that make a video with not only showcase of the project but also try to explain those things which is going to be benefited for you as well as your students. Also again I'm repeating this because I want all of the educators even they are not the part of the live sessions right now. Okay. Because in the closing ceremony of course that would be also live that would be also recorded but that session has to be taken on live because that is only going to make you make it benefited for you because we are not only going to just tell you about how the certificates you are you would be acquiring your certificate and batches but also discussing with you about the I should say about how exactly you can implement this also one more important point which I would like to tell you that you can also be part of the coming closing ceremony. Yes you have heard it right. You can also be a part of our closing ceremony. You can be a part of our panel for that you have to share a small video basically your experience and you have to put it forward in the telegram group as I said you don't need to put it on YouTube you can directly record it with your phone and then put it forward on the telegram group we would be selecting some educators to call them up to join our join our panel on the day of closing ceremony. So this is now I think might have given you let more excitement that you would be of course coming up sitting us with sitting with us in the panel. So don't forget to share your experience recording your videos in the telegram group okay and do not forget to join us on the closing ceremony coming this Sunday till the time you have enough sufficient time to submit your assignment and we are looking forward for wonderful assignment and wonderful experience video with all of you just share it we would be coming to you and of course we will be letting you know how you would be joining us on the on the day of closing ceremony. So that was all for the day from my side. A big thanks to all the educators for bringing and taking out their wonderful time and learning these things with we are of course looking forward with all the educators educating more than 1000 or 2000 students each okay let me just mention each so that of course what the main objective of the bootcamp is to empower the young minds to empower the educators to empower the young minds and yeah so that was all thank you thank you very much yeah over to you Tina. That was an amazing session today Ayush learning how exactly machines are working and the machines are learning something that's quite interesting hope all our educators enjoyed today's session thank you all of you for joining today's session let's see on Sunday bye bye. Bye bye everyone.