 Very good afternoon to all the educators who are joining us from around the world for the day number four in the international bootcamp on coding artificial intelligence and robotics for the educators. So I again welcome all the educators for joining us here. We'll be shortly starting our session just in two more minutes. In the time let me again call up my colleague Tina. Hello Tina, how are you. I'm doing really well. Thank you. Good afternoon all the educators. I would like to welcome you all on this full day to the international bootcamp program organized by Stempedia and Art Park. We are truly delighted and highly motivated by seeing all your activities on the telegram group. It's highly appreciable. We have a humble request to all the educators to fill our survey form, which is already shared with you at the earliest so that we can support you in all your ways. So now let's empower ourselves to empower the young minds. Yeah, thank you Tina. Thank you for a wonderful again reminder for the educators here educators it's a kind request to all of you to please ensure that you fill up the survey form of course that is for your use of course that will give us an idea what we should bring for you. How can I how can we support you in just making this bootcamp one more step ahead for your, of course, I should say success. So now to start with our today's part, as you know, the first and a very important part that we always want you that still if you have not joined us till now on the telegram group so please feel free to join us. Okay, that's great Ravinder, Ravinder Singh sir, or if you have already filled that's really amazing. So till the time you can just let us know dear educators from exactly which location you have joined us in the chat. So yeah, I would again, our request is still the educators have not joined the telegram group please join us using this link as using this QR code and the link would be also shared to you on the chat shortly. Very good evening Abu Edu Life Creators, Benoy, Varsia Ma'am, very good evening, very good afternoon to all and everyone, Mr. Gautam Kumar. Is it Tronyx? It's really amazing to see you all here joined us join us live. Very good evening to all educators. Now, I would request Tina to please give us an idea of how exactly we are making this bootcamp successful with the help of so many collaborators. Yeah, over to you Tina. Thank you Aiyush. International Bootcamp Program is organized by Stempedia and Art Park, AI and Robotics Technology Park by IISC Bangalore. We have been supported by Niti Aiyog, that is Airtel Innovation Mission, which is a part of Central Government's education system in India. It's been our pleasure to have all India Educators Forum and American Indian Foundation to be as our community partner. We also have many international partners like Purpose Smart Education Palestine, National Institute of Education Singapore at the Excoding Academy, Egypt, Informatics Tunisian Association, Tunisia and EduStream UAE. Now, it's time to talk about the basic objective behind this bootcamp program. The objective of International Bootcamp Program is to empower the educator by deepening their knowledge on coding, artificial intelligence and robotics and thus by enhancing them with all the necessary skills that they required on this 21st century. By International Bootcamp Program will give all the educators a hands-on experience on the latest technology like robotics, artificial intelligence, machine learning in a fun and very impractical way. By learning more about artificial intelligence, robotics, coding and all, the educators will become more confident to talk about and discuss more about artificial intelligence and all. By this program, all the educators, we can make you as an ambassador of your institution also. Now, we can see some wonderful creations of our educators. Over to you Ayush. Yeah, so thank you Tina for again giving us an idea on the objective and as I said with the wonderful efforts by all the educators around the world by making the amazing videos with their explanations and hands-on activities. Definitely they are going to be the ambassadors of AI learning in their institutions. So now coming to the part of the wonderful creations. So as I said, we have of course taken out some of the projects by all of you. Of course, there are many wonderful projects. Of course, that would be getting checked by our trainers and our educators. But here I would like to just show you some of the projects which were like amazing to see and we were very glad to see how exactly the educators are putting their so much effort for getting the hands-on on the activities which are getting taught during the session. Apart from it, you can also see it's not only the things which we have discussed in the session, but there were some of the activities like airdrop, which was again a DIY activity when there was an object detection day where we wanted you to make of course the biodegradable and non-biodegradable waste classifier. So it was really amazing to see that how exactly you are coming up. As I said, these are just some pictures to keep you motivated, but there are many, many projects and we are looking on each and every video which you are sharing. And it's as I said, it's always amazing to see the way you guys are explaining the concept which you are learning in the sessions. And that is going to help you as well as all the students which are connected to you. As you already know, our motto is empower yourself to empower the young minds. Of course, we are looking forward with you all to empower all the young minds connected with you. Now coming back to a small notice again, which was just shared by Tina. So I'll just like to show you the form actually, which we want you to fill. So this is a survey form and this is a survey form for all the educators, whether they are from any of the part of the world and any part of the India. So, yeah, so basically here how exactly it has to be filled. It's very, very simple just it's we want you to share your experience with us. Let me just zoom it for everyone. So basically the first is of course the full name. So you would be filling up the full name, share your email address, select the country from the dropdown mention about the institute name. Now, this question is very important that do you have any makers to space stem steam lab, tinkering lab or any still development lab for your students at a school and organization. As I said these questions or you can say the survey is just for your benefit where we would be trying to get something for you that you can implement at your end and of course may give the benefits what you are learning. To the students who are connected to you. Now of course, the next question is which kind of hardware kids devices, do you have at your school. So this will again give us an idea. What are the things we can bring it to you, of course making it more successful journey for your students. Which coding platform software do you have as a part of your training programs if you are having any training programs going on in your school. Or if you wish to have a training program, which kind of coding platform would you like to use now which it devices do you have at your school organization. So basically this includes it devices like laptops tablets desktop iPad Chromebook. So, as you already know picture blocks is very much compatible with almost all the devices still we want to know if there would be anything else which we can help you with. So we want you to give a give us an idea of an IT infrastructure at your lab. Now this is about that how many students are getting a coding AI robotics training in your school and organization, again to get a capacity. So that of course we can help you to implement the programs at your end and similar questions do you have any specific curriculum. And then are you using any book of coding or AI robotics. Do you update your curriculum or resource like hardware or hardware kits or coding platform from time to time. Are you satisfied with the current resources available at your organization and this is a really very important question, because this is going to give us an idea that how exactly you want it to be upgraded. Okay, so that we can try and we can bring our best services for all of you. And of course at the last if you have anything to be added anything to be mentioned, please feel free to put it here, but it's really important with by every educator it has to be filled. Now, I'll just share my screen back. So, as we were discussing about the wonderful creations, and with this wonderful creation, let me just again give you an idea as we are approaching towards the, I should say there are only one, there are only two sessions left in this boot camp. So we want now you all to start submitting your assignment, although we are getting all the assignments from you, but still, if you are lagging behind please take your time and start submitting your assignment, because once you have completed the assignment, of course you are going to get your matches in the certificate accredited by artparkstempedia.org. Apart from it, just be patient with us till the end of the boot camp, each and every session is of course going to help you to enhance your knowledge, and let me tell you towards the end on the last session of course, we are going to not only enhance your knowledge, but we are also going to help you with the execution and implementation that how you can become an AI ambassador of your institution, where we would be taking talking about helping you to get the AI and robotics teaching resources for yourself, as well as the access of the students learning resources. So we are looking forward with all of you to be part of the boot camp till the end. Okay. Now, I would request Tina to please take us through the schedule of the complete boot camp and where exactly what the things we have covered, what the today's topic and what are the new things coming in the way. We have two boot camps that is level one is coding and artificial intelligence. In that we learned about interaction to artificial intelligence that was session one in session two object detection in session three, we learned about human body detection and did a hand control game. And today in session four we'll be learning about natural learning processing will learn more about text classifier how how different home automation, home automation techniques, and we'll do an activity in session five will be learning about machine learning how exactly a machine learn something will do will understand how to make models in machine by using machine learning, and we'll do an activity of animal classifier and the last session is our closing ceremony. Yeah, so I would like to just add an important point here so be at all educators make sure that you also see the days. So today is of course the 23rd of June. Tomorrow is 24th of June where we would be discussing about the machine learning, and then we are going to give you a rest of a one day and then on back on 26 June as I said that would be and going to an amazing closing ceremony where we are trying to get the wonderful experts from from our team to get you know how you can of course implement the thing. And of course we are going to talk about that how exactly you are going to get the resources and many more exciting things. Yeah, so moving ahead with the book, the schedule of the book come number two. Yeah, over to your Tina. In level to that is robotics and artificial intelligence heroes of we have six sessions session one is introduction to robotics will be teaching you the basics of robotics, and we'll make our own robot to move in the second session is self driving robot will understand how sign detections using artificial intelligence works and will understand the concept of self driving car and will make our own robot, which is controlled by science. The next session is session three that is line follower robot will understand the basic concept behind the line follower robot and will make our own robot to follow a particular path. The fourth session is to make our robot as a delivery robot that is AI delivery robot and will learn the logic and will understand different application regarding the same in the fifth session will learn about the guest control robot will understand how poses are identified using machine learning and will make our robot to move by use by identifying our guests and finally the sixth session is closing ceremony. Now, are you sure I would like you to take forward to discuss more about how we are going to move ahead by learning through this live session. Yeah sure Tina thanks for giving us a good update and overview of the schedule I think educators might be very very excited on this part. So now already I think many of educators already know they are submitting the assignment, but still if you are facing any of the issues let me just take you through about how exactly you have to go through with the elements. So let me just click it here on the international book. So basically educators this is the website that is a dot tempida.com. So if you have got your username and password of course you can and directly log in and enroll to the book. Still if you have not enrolled so make sure that you sign up in this website. Okay by clicking on sign up option. And then you have to click on enroll now. Once you have clicked on enroll now you have to click on my courses on the left hand sidebar. Once you will be clicking on my courses you would be basically able to see the boot camp. Now here you have to click on AI and coding boot camp for educators level one. Here again you can see we are still going on with our doubt session, which are going to be from 6 to 7pm and 10 to 11pm ISP. If you want to join of course you can join click on join and join us from here. If in case you want the direct link you can always feel free to check the telegram group. Now coming to the part how exactly your learning process is that as you know we have completed introduction to AI, object detection, human body detection and today we are going to work with natural language processing. In the natural language processing, you can watch of course the right now the video which is live on the YouTube playlist of this Tempedia apart from it you can also watch it in your own dashboard. And here you can see this embedded video where you would be able to see the live session as well as this video would be there for like of course for the next coming two years this LMS is all yours. So this video will be still here. You can see these are the feedback in the attendance form built by the educators from all around the world and who especially joined from the Atal Tinkling Labs and an AIM team. Here you can see this is like a complete content based learning program. So here you can just go ahead with learning through the content shared to you. And at the end you have to click on the activity which has to be submitted by you. You as an activity you have to make a small video of your project where you have to not only just show the working but try maximum to showcase each and everything starting with explaining the concept, working on the court and explaining the code part and of course at the last you can you have to basically show the working of your project. Do not forget to name the assignment with your name underscore the name of name of the activity. So this is the way how you can name the assignment and paste the YouTube video link. Do not forget to keep your videos in public that is very very important then only we would be able to see that. So do not forget to share your YouTube video link here and submit it towards the end. Of course, once you are very well with the concept, you are always welcome and we highly recommend you to try this that it do it yourself activity that is a virtual doctor. So I hope this is everything is clear and now of course we can go ahead with our learning journey for the today with as you know that today's topic is NLP. So I'll just share my screen back. Okay, so I hope my screen would be visible. So you're talking and of course starting with NLP. Many of you might be knowing that what exactly the NLP is, but to start with very basic is complete full form is natural language processing to understand every word. It's like to understand the concept behind the natural language processing, you can just understand it by we are processing the natural language that is the human language. Generally computers and machine are good for working with tables, like you might have already worked with spreadsheet data. If you might have already worked with Excel sheet Google sheets and many things. So of course they are very good in working those data. However, as people communicate more often are in words and sentences. That means generally when we are talking to each other, what happens is we are of course not talking always as an excel sheet or a data we talk in a natural language. So many of the details that people say and write are unorganized way that means how we are speaking to each other. There might be some grammatical mistakes there might be some language barriers, but of course it cannot be as same as the same as to every person. Okay, so that is why what happens is it becomes little typical for the machine to understand this. So but how exactly we can make our machine so smart that they would be able to basically understand the what exactly we are speaking. So as you can see the goal of the natural language processing is basically making the computers understand whatever we are talking it in our languages, whether it can be of any regional languages or any international or national languages. Now, this can be all this can be all done using of course using an idea of artificial intelligence, where we are going to make our machines more intelligent. Okay, where they would be able to not only interact with the humans, but they would be able to understand what exactly we are speaking of we are talking to them. So yeah, again to summarize, this is natural language processing where we are trying to make our machines smart in such a way that they are able to process the natural language on which we are talking. Now, how exactly it works it's not always so easy for machines to understand what we are talking. So here I would like to tell you, there are many algorithms working behind the natural language processing, but to understand it in a very simple way you can break this thing into four different steps. The first thing is, of course, the machine requires to capture what we are saying. Once it is captured then human languages converted to the data that computer can understand. Now, of course, as we know the computer can only understand the zeros and ones, or you can say the mathematics of mathematics part. So there are algorithms, which are going to help to get process the capture data by you by the machine from you to make it understandable for the computer. Now, these natural language algorithm natural language processing algorithms are trained, of course, by the human on already the data, which are like a further data of speaking, like, if you might have heard of Alexa so it is able to understand what we are saying. So somewhere behind the scene, there are different models and algorithms, which have been trained on giving different instructions that if someone says hello, so machine has to reply the hello. Okay, if someone says that I want the weather data. So that means machine should understand that weather data is basically referring to the climatic condition of that particular location. Now, of course, that has been already trained. Still, there are many of the researchers or researchers all around the world who are still working to make this possible that every regional language of the world is recognizable by the machine. Now one of course the data is selected. We have converted that data for a machine understandable format machine has understood it. Now machine is ready to take the new data on which it is going to respond you. And there we call we can call it as a testing data. So when we are talking about the testing data, that means we are giving a new data for a machine to understand it. And the new human language data is given to the algorithm and the machine would analyze that whatever you are saying or the data or the I should say the text format you are giving and at the end will give you the result. Now, in this, of course, NLP if we are talking about, so there can be the multiple ways we can give the data to the machine. There can be a data where we are giving it from our keyboard. So just think of the speeches. Okay, so if I would be talking about there are speeches written in a word format and that data is given to the machine for understanding it. So of course, machine would be able to get you the analysis on the data. But to make the machines little more smarter and make the humans to do little less serious work that if how it could be much more useful that if we make the machines directly understand our voice or our speech rather than put inputting the data in the form of words. There comes the part of speech recognition. I know we already know many of the examples of the speech recognition starting with Alexa Siri Google Assistant. But what exactly does the speech recognition refer to? So basically speech recognition is the ability to machine to identify the words and the phrases in the spoken language and convert them to the machine readable format. Now I'll tell you here. Yeah, so coming back to the part of speech recognition. So yeah, so basically if you'll try to relate what we have understood as a part of natural language processing. So here I'll just go back to the slide and want to drag your attention on the parts of first human languages converted to the data that some people can understand. While we come into here, you can see, identify the words and phrases in the spoken language and convert them to the machine readable format. Now here we have added one more step. Previously, what was happening that data was ready in the form of text. Okay. And we go ahead, we go, we give that data to the machine directly to understand. But when we are talking about speech recognition, then a thing comes up here where our speech is getting recognized our voice is getting recognized. So in that part it becomes little more typical for machine to understand what exactly the word I'm pronouncing. Okay, like you can see there are different ways to say hello. So we can say hello, namaste, hello, and many other in many other languages. So it becomes little more challenging for the machines, but of course humans are very smart that they can of course train the machine to do those things. Now, when we're talking about speech recognition, okay, so when we're talking about the speech recognition, and you can just relate this as I said relate this application to any of the assistants you are, you might be using and if you have not used, of course you should minimum give a try to that to get an experience of speech recognition. Again, there are many algorithms working behind it, but let's break down into the three major states. The first is the machine has to record the audio file because machine cannot directly complete directly convert what you are saying. Of course, our brains are very powerful. Even if I make a mistake in my communication, you would be able to understand what I'm trying to convey. Because of course there are many languages around the world, but till we are able to manage to understand even we don't know that language. But for machine it's not so easy. So that is why they record the file. Although that recording is totally upon the machine or the makers that they generally deleted just after converting the thing. So the first thing is the recording the audio file. Now, the recorded file has to be processed and in the process basically they break down that audio to the text. Like right now if you're going to record the file, whatever I'm speaking, then is not any person but of course the machine is doing that that they're writing down the text what has been spoken by me. Of course, the older things would not be recognizable by the machine, but they will try their level best to come up with a particular text format of the audio. Now, for machine to be intelligent or a human to be intelligent, it's really necessary that we have a good knowledge base. If we are saying that of course as a human I am like of course intelligent person that means I should have a good knowledge base I should be able to speak up or able to understand the different topics or a different thing, whatever is going in front of me. Only if you are saying the machines are intelligent, then of course they should have a knowledge base of the word that is if you are talking about English, so they should have a knowledge base of English word. And that database is going to help the machine to identify the word what was on the text, which was taken out from the speech, what you have given. And of course at the end, they will try to understand it as they will try to make it make it usable and understandable by the machine, and then machine is going to come up with the solution and answer to the things what we have asked. Now, this is a small you can say pictorial format of how exactly the speech recognition work. So I would like to know basically from the educators side that in their opinion, what would be the step number one. So I would request kindly to all the educators, if they can write on the text, what would be the step number one out of these four steps that is processing responding speaking and listening. So can we have some answers on the chart. Until the time you give me the answers, I'll just get ready. My picture blocks are going ahead with the activity. So that's superb. I have started getting the answers that is speaking, speaking, listening, okay, listening, speaking. So I think you might be little getting confused between responding and speaking. So you have to understand, responding is something which you are getting out of the machine, and speaking is something which you are giving out as a user. So yeah, of course, the first step would be speaking, because till the time we are not going to give the data machine would never be able to understand it. So that is why we have to start with speaking, we are going to speak up the things. The machine is going to listen. That means machine is somewhere recording that file. Now comes the processing part that we have already discussed that in processing, of course, those audio files would be broken down into the text format. So using the text format of the word, it would be getting compared with the knowledge base of the machine that what exactly that word means. And after understanding the concept, the next part would be the responding. So, I think that is for sure. So, thanks everyone, thanks Binoy, thanks Ravinder Singh, Anup, Mukesh, Shwetaji and everyone for responding on the chart. Thanks everyone. So now, coming to the next part that we have understood the concept. Now we have to make an activity as well. So now we'll start with the activity. I hope you all are ready to make the activity. And as you know, like today's activity is related to making a text classifier. So let's start making a text classifier in the pictoblock. So now I would be sharing my pictoblock system. Okay, thanks Anamika, thanks Shilpa, Maninder Singh for responding on the chart. Yeah, so you're back to the activity. I'm sorry, I'm a little low with my cold and cough, but okay, not a problem. So coming back to the activity part. So here, as we know, we have to work with natural language processing. So in the natural language processing, of course, we have a separate extension, which is going to help us to learn more on it, do more hands on activity. So what I'm going to do here is I'm going to add an extension from our extension option. As soon as I click on add extension, you would be able to see the extension that is natural language processing. Now let me just click it here. And let me just show it to you the different, I should say, blocks related to the NLP. Now, here basically, you can see we have a three in the text. Let me just zoom this. Yeah, we have first block as add your text as class. So basically, we can add a certain text to a class. And when we are talking about artificial intelligence, a class refers to the label of anything, whether it's an object detection. So we can call a car to a different set of cars all around the world. Okay, so if we take a class, if we take a car as an class, so there can be all the cars in that particular class, like we can take many of Maruti and any of the brands cars, right? So basically, we are labeling all the cars with different brands with just tagging them as a class. Similarly, if we are doing an object detection. So of course, as I said, like bicycles, so all the bicycle would come under the bicycles. If we're talking about cell phones, all the cell phones would come under the cell phone class. Similarly, when we are doing it in text classifier, so we would be making a small class. Okay, and we would be adding more and more text to it. So that we are a machine should be able to know that whether any of the things have been told the meaning is same. You have to understand this, like if I say turn off the light. That means I want the final action to be that light should get switched on. I can say it in a different manner, switch on the light. I can say I want brightness and many more. But at the end, if you collectively find that we want only one word, that is switch on the light. Okay, or you can just cut it down to on the light. Okay, so basically, this is the class. Now the block says train the text classifier. That means basically we have to train here. It's not we are using already trained model. So basically in some back end, we are going to train the model. Now the block says reset the text classifier, because there can be many times that we want to work with a different set of things. We can train our classifier on switching on the light and switching off the light. Apart from it, we can also get it trained for moving the object. We can also get it trained to turning on the different appliances. So for that basically we can reset the class to reset the text classifier. Now, apart from it, here you will see we have get the class of your text. So again, once you have trained the machine, okay, I'll tell you what exactly is the beauty of natural language processing. So basically, you are going to train the model and then you are going to put up a text which can be the part of the training, but we would recommend not to use the training thing. Like if in case I'm training my model on switching on the light and I have given four instructions, switch on the light, turn on the light and many more. So but I have to use basically any other statement. And then if my machine is able to understand that then only we are going to call it intelligence. It's similar to the way that if I if we have been taught or if we are teaching the students that to multiply by two is four to multiplied by three is six. So we want them to come up with two multiplied by four. Okay, so it should not be that we have taught them to cross three is six. And then we are again asking what is to cross three, that is not going to exactly make them intelligent. Intelligence would be somewhere coming when they are able to take the decision after learning the concept. So I hope this is these things are particularly clear. Apart from it, we have a get a sentiment, a sentiment block, which is can help you to get the sentiment, whether it's a positive or negative, I am have very happy. Let me just delete some of the part I'm very sad. So it is going to say it's a negative. So it can take in all the lies it's like one liner, of course, but it can have two to three things. I am sad. But I was glad to meet you, meet you guys. Now, I am happy to share the knowledge with you on. Okay, so if I have written this if I'll click it here, you can say sing neutral. Why we have used a sad as well we have used a glad as well and we are also mentioning happy. So this is how exactly it can let you know about whether what. Okay, so basically if I say that I am very sad, but I was glad and now I'm very happy with saying it neutral, because I have I was one part of my speech was I was sad. Some part of my speech was I was glad. Right. So there can be three types of sentiment, first is positive, negative and neutral. Okay, so you can have a complete speech here inside it and of course check whether what was the sentiment of exact using of those words. So now I think you would be very clear with the set of blocks in it. Now we can start with the activity. So now the first thing comes up that we know we have to use when green plug is clicked as to start the shifting part. Now, if I asked you that please make a project on, let's say natural language processing only okay, so just make a project in natural language processing. You, some of you already know it you can make it, but some of you would say you have not thought of anything you have not shared your knowledge, you have not trained us basically right. So how exactly we are going to work in a similar fashion, we have to also train the machine first, and that is really very important. So now what I'm going to do is I would use a block which is going to train the data. Okay, so basically I have to train the data. Okay, you have to understand how exactly I'm going to train the data and then we are going to use it. Okay. So here I'm going to use make my block, sorry my block option. Okay. So when I selected my block, what exactly is my block, let me first tell you my block is basically a way in pictoblock to give you a functionality of using the same code again and again, without even using those set of blocks, even what you can do is you can match up all set of blocks with one block, and then refer a small block that would be calling up all and all complete different set of blocks that you can if you want if you are already little familiar with the programming. You would be able to see, you would be able to see basically functions in the real programming. So this is like a my function thing. Okay. So now let's first let me click on make a block. I want to name it as a train data. So basically I would be training the data inside it. Okay, so I'll click it here. So now I've got this train data. So our first what I'm doing is I'm making a function. Okay, to train to keep ready the train thing. Okay. The first thing what we have to do now, as I said, you are using the same pictoblocks again and again. Okay. So there might be a chance you might have used the NLP extension for certain different purposes before. Okay, now you want to use it again. The training of this data is happening offline to the pictoblock. So it is saving all the calculation inside it. So do not forget to first use the reset the text classifier that if the class you have only one reset or one classifier inside the pictoblock. If you have trained it for saying something else, it is going to do that. But if you wanted to work to do anything else, in that case, you have to first clear this down. And that is why we are using a reset classifier. Now, of course, we are we have to basically not train the data. Okay, now, what type of data we require. So here basically we want to make a project where we can turn on the light or turn off the light. Okay, I know. Can you increase the font size? Okay, I'll try that. I think now it would be visible from keeping this train data aside as of now. Okay, now I have to basically first create a data. Okay, so as you know, in artificial intelligence, the first important part is acquiring the data. The second part is training the data. The third part is testing the data. And the fourth part is of course deploying that for making a useful application. So the first thing we have to do is we have to train the data. Okay, so let's train. Let's collect the data actually then we are going to train it. Okay, so I'll make one more block here, which I'll give a name as data for light on. Okay, I'll make it here. So now basically what I'm doing here is I'm collecting the data. Okay. Kindly elaborate train and how it works at the back end. So basically, at the back end, there are many different models and the algorithms which are working behind like a bird. One of an famous algorithm is bird. What happens, we can of course discuss that I'll tell you that is a very long process where what happens is there are algorithms. They basically cross verify what exactly the data you have given and then using a bag of words and different concepts, they collectively come up with a final training part. And once they are trained, then they take on the data for you and of course reply you. So if you want to explore on that of course I can share you some of the algorithm means one of it is bird. So you can use that and you can of course look into it how exactly it is working. Now, we are collecting a data here. So I think you are already familiar till now with this block with stays. Let me just bring other blocks down so that you don't get confused. Okay. So now coming to the first part that is data for lights on. So now here what is my motive is that I have to make a data first. Okay. So now exactly what I can do is I can write here the new data like I can write here live on. Okay, now I have to also define a class and that would be basically will remain same for all the data. So I can just to make it simple we can keep it light on. Okay. Now I would add the next so I can just duplicate this and add it here. Okay, now I only need to change the text part. I don't have to change the class because the class will remain the same. It's too dark here. Okay, so I want that even if I'm responding my machine that is too dark here my machine should be able to understand. Yeah, this is somewhere my user wants the machine to switch on the light. Now again duplicate this and put it here. Now I want if you can please suggest me some of the phrases you want me to keep here on lights on that would be actually amazing. So I can write I want to study till the time you are sharing your phrases. Okay, I want to study with the uppercase and that yeah of course it's not going to make any such big difference because as I said it's only working at the back end with the meaning of the world. Okay, so it's not going to make so much of difference, but it is going to make a difference on the class part. Like if you have given class as lights on it should be same. If you have given the class name or lights on with very small so you have to maintain that because you cannot change the class name as it's a label. Okay. So because even if we know that there are like light on light on lights on can be or I should say would be treated as a different classes. Okay, okay, it's light. Okay, of course, we can take that. Okay, shine up. Okay, that's pretty cool. Okay, let's sign up the room. Okay, I'll add this shine up the room. Okay, let me add some from my side. Turn on the light that is generally we say it dinner time okay. Okay, let me know some more light is too low, we can keep it light is too low. Okay, clear the dark. Okay, that's pretty like creative. Okay. Okay, now we have to also keep it that we are training it. Okay, we have not to make it a little more complicated for the machine. Okay, so we can now have a normal on as well on that also means a switching it on. I want to take Shah. Okay. Okay, so now I can I would like to add one more. If I can get it from turn on the electricity or first turn on the light is already there. It's already getting dark. So we have already used it's too dark. So what we have to try here is that we are using a multiple things here. Oh, I'm sorry. We are using we are trying to give the input in a multiple way. It's too dark here that of course there is a dark so anywhere the dark would come the machine would understand studies later is now done shine up the room is done light is done. Okay. We can mention one light is to dim. Switch on the light of course switch on the light. Let me just see we can keep that of course. Okay, so can we change the language. Okay, yeah of course you can change the language but it's not going to help you right now because right now natural language processing is not going to understand which I like her though because the right now the natural language processing which we are using is going to train only on the English language. Okay, so I'll just add the last thing switch on the light. Now, okay, I think we have a pretty good data let's let's count it actually because we have to keep the data even even fit the both the classes. So this is 12345678 and it is sufficient. Now let's go ahead and make one more block that is going to help me to data to turn off the light so let me just write it now and name as data for lights off. Okay, and I think that's ready and let me just bring it down here. Okay, so we have actually now sufficient data for the keeping it on. So now let's train it to lights off. Okay, so. Again, I'm sorry it's not we are training right now we are just gathering the data. Okay, so let me just again remind you right now we are not training we are just gathering the data. Okay, so now the next class we want it should be lights off and even we can keep it like big for that like we were comfortable with it. Yeah. Now let's lights off. Of course, the first is we can take lights off directly. Now let's. Yeah, let me just duplicate now I want you help me to get the things related to switching off the light. So I'll just use it to write here and your lights off would remain the same. Let me just add a block here and now I want some suggestions from the educator side. Okay, switch off the light of first that's perfect. Okay, so how do I upload it to Arduino. Now actually you won't be able to upload this to Arduino since this requires an Internet and Internet cannot be directly exist inside the Arduino. Okay, so you won't be able to, but you can try to implement some things where you can take input as a data and put it here through the Arduino. It's time to sleep that's perfect. And let me add some more blocks here on which we require. Okay, six, seven and one more. Okay, so turn off the light. I think we have not taken it. Okay, let's take that. Turn off the light. Okay, okay, that's actually would not help actually, because it would be little like rude if we in the home we are going to save electricity so I think that yeah, I want to take rest we can keep it of course I want to take rest. Okay, I do not want to study. What do you think which can we take of course we should take because as I said the data should not so should not be so simple even. Okay, so I do not want to study. Like, since I wanted to study so I wanted it to sorry. I do not want to study. Okay, now some let's one know one more we can take is off. Okay. Yeah, let me just keep it off. Okay. So now I think let me just give you a small sorry small idea what we have done I know this would be very small visible under screen. So I want to give you an and make you understand what we have done. So if you will see I have given the data for lights on I have given here the data for lights off. Now we are going to come up to make a function where we want these things to happen or this data to be utilized to train the machine. Okay. So we what we did of course we reset it the classifier. Now what we want. Okay, I want these data to be called in. Okay, I want that my machine should go inside that data, I will bring lights on and lights off block. This is what I was trying to tell you. So now see if I want to again use that I want to use these data. So I don't want to like write all these things again. This is why what I have used is my block where I have defined up thing. I have defined the thing and now wherever I would require it to be used I can directly use with a single block I don't want them to be repeated and written it again and again. Because of course that is going to make the strip clumsy and little difficult to understand. Now of course at the end I want the training path and that is where which is very, very important. So now we are ready with the training as well. Let me just bring this green flag here and let me just zoom it here to show you up the thing. Okay, now till the time what we have done if you will see I have defined here something I have defined here something in the second part that is defined data again I have defined this. Okay, but now I have to call it. Okay, I have to call these things. So now just with I think there is a small mistake. Let me just check. Actually I'm not able to get one block. So let me just define one more block here. Okay, now here we go. So due to some glitch I think there was a problem that I was not able to get the reporter block of this definition. Okay, but I'm now able to get it. Okay, so now what I'm going to do is I'm writing this script just very simple that when I press a green flag I want to call up a train data. So what is going to happen here is let me just bring down my pin. Okay, so this train data is basically going to call this defined my block. Okay, so this is going to call this. Now, it is going to work blockwise block. So first it is going to reset it. Now, as soon as it is going to say data for lights on it is going to somewhere to this data from starting of course. Once this is done, then this particular function would be called or I should say my block would be called and this my block is going to call up this particular data which is for all. And at the end, after collecting this after collecting this it is going to train the text classifier. Okay, so I'll just clear up this. Okay, now, what I want. So this is what I made. Okay, but of course I want at the end something right. So if you'll see, let me just bring the data and just show it to you again. So I have here written I want to study. Okay, and even even here written I do not want to study. Now if in case I want some input to be given. So right now first I'll show you how easily you can do it. So you are going to get the class of any of the data. So you can see right now anything is written and that is it is saying lights off because somewhere it was previously trained till now it has not been trained on the new data. Okay, so now I would add it here something like need to study. Okay, need to study. Do you think what it should come up lights on still if I have not been let me just press this green flag. Now it train each and everything. And now if I press it, it's saying lights on. Okay, now if in case I would need to rest. Okay, if I click it here, you can see it showing lights off. I think it's very, very small. Let me just zoom it to you so that everyone would be particularly able to see the output. Okay, so here basically what I'm doing it if I'm saying need to rest. So it is automatically showing me lights off. If I'm here writing just I'm changing word word. So that is I think we're making. Now you understand that it is making the center not the complete taking up the sentences but it is working on small words. What exactly that means. So if I'm pressing need to study it is saying this. But of course, this is class is not going to help me anyway right now. Okay, so what we can do is, okay, I can ask my toby to say this class, I can go to look and use this. Get the class off. Okay, now anything has to be given here. Okay, anything has to be given here would be giving you an answer lights on or lights off. Right. But now do we need to write here, because we are making a program of course we do not want it to be like this. So again we have a sensing palette. Okay, and here basically we have a block says ask a question. So let's why not use this. I'll put a ask a question and we will make frame up a question what you want me to do. Okay, and we can write comma, or we can use a question mark. Okay, and instead of this we can also use directly the answer. Okay, we can put this here. Now if I would be clicking it here, it is what we want me to do. I want to study. And I'll put it here. Yes. And lights off. Oh, I'm sorry. Let me just check. What's inside this. I want to study. Okay. Let me just check this need to study. Okay. I want to study. Let me write this. Okay, so see right now it's not working good with I want to study that means I need some more training on my data. Okay, so I want some more because there might be it's not able to understand what exactly I want to say with I want to study, although I have given it here. Just let me bring this down. So let's why not try the same thing what we have that is need to study. Okay, so let me just try to use this need to study. Okay, and if I press it here. Okay, same lights on. Classifier is not trained. Yeah, it's not sufficiently trained you can say if it is working for something it's not working for something. Of course, it requires little more training, because the more you would be training it on the more answer, the more well it would be behaving right, but still, if you want the things to happen as well. And you can add the things here. Like, if you want to see you right now what we did we are asking a question, then we are inputting it. Now, even what you can do is to make this project more beautiful. What you can do is you can directly use some external, I should say, you can use some external right of well. And you can keep the two costumes, like one is with the switched on light and one with a switched off light. And then you can ask what you want me to do. And if a particular answer gives you or the particular thing gives you a light on, you can change the costume. Okay, so this is how exactly you can add the things to here. Okay, so now if I want to try I would require actually some more Sprite. So, let's try. Of course, I can try this. Let me just bring this button. Okay. Okay, and I would try basically to first get a green flag here for it. And of course, this is very pretty small. So, I would go to look to increase the size of it so that it's particularly visible or not 2000 that would be a little bit. Okay, now this is good. And now what I want initially I want its costume to be. Let's, we have one costume here. Yeah, we can keep it like this. Okay. Now, you can use this to turn off and turn on the light. Okay, you can change the costume. Now how you can come from here to there, you know, I think there is a concept of broadcast message. Okay, so what you can do here is you can use two two conditions, that is, if and if. Okay, we are going to use two conditions if bring up the operator. Okay, I'm doing it parallely to make you understand what's the difference between it. Okay, so I'm using it to operators, particularly same. Now, if what is exactly instead of say, of course say let it be so we can take the same block again actually if ticket and whatever you have entered, if the class is equal to. There can be the two different classes. There can be only two different classes which we have given lights on or lights off right. So, here basically you can make conditions, and you have to make sure that while comparing this, you are very good on checking that what exactly the uppercase in the lower case, because when you are comparing it it is going to compare it with lower case and uppercase. Okay, and duplicate it here and put it here. Now we can use this. And we can broadcast the message. And we can come it here. So let's not make it for two things. Let's just make it for if an is okay. So if lights on, I want a message to be broadcasted. Yeah, so basically I would be broadcasting a message here. And no say message here would be lights on anything would you want to keep lights on. And I'll go here. And I would use a block. When I receive light on what I want is. I want to change the costume. So right now we have initiated it with button for B I can take it button for a here. Okay, now you just you will see what you want me to do I need to study. Okay, I need to study and as soon as I click it here lights on a safe it broadcasted the message and the light got turned on. Okay. Classifier is not in here as I mentioned you can add more data. So now I think you can see it here. Okay. Similarly, you can add one more block here similar to this where you can check light off. Okay, so let me just duplicate the same thing. I have to just bring it out. And instead of this we can keep it off. Okay, and this time I want a new message to be broadcasted that is light off. Okay. And now the same thing I have to do it that when I receive a light off message what I want I again want the costume to be switched back to the 4b. Okay, so now if I press it here now every time so right now you can see it switch off. So if I say switch off. Okay. No sorry need to study. And if I say yes, so it is going to check lights on and it is going to bring up lights on. Okay, but if in case I say it need to rest. So now if I press it, it is saying lights off and the light is not going to off because we have initially put up the lights off. So even you can keep it like this to see the changes where we do not specify anything on the initial. So need to study. And if I say this it is going to basically turn on the light. And now press it again and write need to rest. Okay, and if I now press it here, so it is going to turn off the light off. Okay, bit. No please can you please repeat from the news page. Sure, I won't going to repeat that. So now I'll tell you. So, till here I think everything is clear to everyone. Okay, that we train the data we got the data. Now to depict what we wanted the end is it should also look good that some light is getting turned on. Okay. So what I'm using, I'm using a sprite. The sprite name is the button. Now this is right has to costume. So what is costume person is understand this the costumes are nothing, but the two different pictures of the same is right. If I click it here on Toby, you will be able to see different costumes here like walking, walking to walking three and many others. Similarly button also have a costume that is one is green and one is black. So we can use it for our own purpose by changing its costume. Okay, so I will just go to block it. Now I have added the things here. Now I'm just comparing the two things. Okay, that is, let me just zoom it here. That is what I'm going to do here is I'm comparing the answer. So I'll just break this condition and show you how exactly it is working. Okay, so these are three blocks. First is get class of your text. So for which text I want the class whatever I'm putting the answer. I think this is pretty clear. Now I want this particular thing that is what would be the class should be compared to a word. Okay, that's what exactly that class is whether it's the lights on whether it's the lights off. So I'm comparing it using the operators. The operators are the blocks which consist of all the blocks, okay, which are mathematically can help pull for you. So basically what is going to happen now I'm going to compare what exactly inside it. So let me just show you it has lights. I'm sorry, it has lights off inside this block. Okay, I'm talking about only this block right now. It has basically lights off inside it. So if I'm going to compare it to lights on the condition is going to not to be followed or I should say condition condition will get to false and this particular block will not be executed will not get executed because this condition is getting false. So now I would add all of these things. Okay, I'm going to compare it. If this condition is true, then only in that case I want to broadcast the message. Now you would be thinking what exactly is broadcast message. So when we want to convey any message from one sprite to another, we don't have any other way, leaving to broadcast a message. That means we can message like how do we use broadcast message in WhatsApp that if you want to send one message to everyone. We just write it and we broadcast it to many people. Similarly, from one sprite, you can broadcast a message to all other sprite. Now, we discussed yesterday about cloning. So what was the part is, we discussed about cloning though, if I am creating a clone, I have to use when I start a clone. Similarly, if I am using a broadcast message, okay, if I'm broadcasting any message, so I have to receive that message to if it won't be received by anyone then what is the use of sending the message. So that is what I'm broadcasting a message in new message. It's a text area. You can write anything as per your convenience. So I've made a condition and I kept it like if this condition gets true, please broadcast a message to everyone that we have certain true things with us in the sprite of button. First is, of course, I increase the size, okay, because the button was very small. And I mentioned that if I receive a message that there are only two types of messages, which would be sent out of the first sprite, that is either that is light on a second light off. If you receive light on message, then you have to change the costume. And I think you can just see the name of the costume here that this costume name is button for B and this is button for button for a. Okay, so that is what I changed that I kept it here. Let me just zoom it here that I receive lights on I want button for a if I receive lights off I want button for B. And now at the last it was working so of course you can add many things you can add your own sprite here change the color change the things and many other things. But I think this is particularly would be very clear to all of you. So now I want basically everyone like if you have any question, but I'll just take you back to the presentation part here. I wanted it's very important to discuss the application of every concept. So when we're talking about application of a speech recognition, basically, in the speech recognition, it is getting used in many of the areas. As you know, I should say, I'm sorry, I'll just reshare my speech. Yeah. So if I talking about the applications of speech recognition first. So the chatbot. So this is particularly like a chatbot what we need right, we just sent a chat that I want to switch on the light and light what is the personal assistant like Siri Alexa Google assistant and more medical dictation software. So you have to understand, as we know in the medical it's really very difficult sometimes to understand if a doctor is dictating something. Okay, even in something it is getting difficult for in the law profession. So nowadays, there are different softwares getting made where it is taking the dictation and getting you the things written automatically, helping the physically challenging people to access the thing. So now you can think just as if they are physically challenge they can just save the things and they would be able to get the access of whatever the things like switch off the light switch on the light. Similarly, talking about the applications of NLP. One of the best application of NLP is auto correct and auto complete. If you have might have already seen you write the things and automatically it let you know what could be the correct format and what could be the correct spelling of it. Social media monitoring. This is like a sentiment analysis that automatically whatever the post is posted it get detected whether it is what can be the sentiments or what can be whether it's good for the society or not. Survey analysis hiring and recruitment. Nowadays hiring and recruitment is also getting taking the use of it where you can you can just check an application of ETS basically automatically it is tracking the application and it is with your resume it is able to track. What are the things you know what are the things you don't know and automatically calculate the resume that whether it's eligible candidate for the interview or not. Email printing we already know that inbox spam update is automatically get segregated through the NLP. Now we have one more wonderful use case like do it yourself which we want you to do is the virtual doctor. Okay, so now the next part is if you have not joined us till now in our telegram group. Since not followed the same sir it's not working I followed the same stuff so I can share you the link okay so that of course you can try but we want you to now try new things as well okay so please try the thing from your end as well although I will share you the link of this file. But I want as I said he it would be great if you would be able to come up with new slides, although you can use the core not a problem. So yeah, so I would request Tina to please share the link on the chat. Yeah, coming back to the part if you are not still a part of the telegram teacher community please be part of it so that of course we can support you. We can help you we can support you with the issues and the queries of your okay, and sorry, a very important we want you all to fill the feedback form. So, as we already know that we have to separate forms basically we're for all the educators who are from all around the world in India, they have to fill the first feedback form. The second is basically for the educators who have joined us through the atl community. So please feel free to feel free to share your feedback and attendance. And of course we are looking to your feedback and attendance daily, and we're trying our level best to bring the best of the best what we have with us. So I hope you would be filling up these feedback form as well as importantly you have to fill the survey form. And as you know, we still have the doubt sessions going on you can always feel free to join us okay on the doubt session at 6pm to 7pm. And of course you can join us on 10pm to 11pm. So could you please give a revision as you study. Okay, sure I would. We have two minutes. Okay, so I will try to give you a small revision of what we did. Okay. Okay, so let me just go back to the first part. Okay, let me just so we were working on NLP so that is natural language processing. So complete your assignments till 30th okay we can extend the date but we want you to submit ASAP the more you would be delaying the assignment submission, the more delay the there would be delay in certification and badges distribution. So we want all of you to do it as ASAP okay. So coming to the revision part of course it's necessary, although I would try to keep this revision tomorrow as well to give you right now a small revision. And NLP that is natural language processing where we are using an idea of converting the natural language to be processed by the machine to understand what exactly humans want to convey. So what we did here, the first part of our activity was we collected the data. After collecting the data, we then trained it. So we have to train the machine make them understand that yeah, these type of data are classified with certain name of the class. These type of data are certified with certain name of the class and that is what we did in this part. The next was we use the functions that is trained data to accumulate all the data reset the classifier if it has been trained previously on certain things. We train on it. Okay, and then we use a small if conditions and a broadcast messages to share the message to the different is right. We are we are we have added a new sprite to just get in. You can say a physical to get an idea how exactly it is going to look like a switching on a light or switching up a light. Basically, the major concept was NLP and I hope that is pretty clear. But still, if you have as I said, you always feel free to come up on the doubt session. Definitely me and my team would help you to understand it much more better way. So, yeah, so now our doubt session. So if you have any doubt, please, please do clear the last part of today's session. So as I said, as we just discussed that we had broadcasted a message, you have to understand, we are sending a message. Okay, two different is right. Different is right sir different. You can say the different characters right so in a movie if you'll see there is an actor, there is an actress. Okay, so of course we are human if actor completed complete is his part. Of course actress can come into play and come up with their own dialogue, but just think that they are unknown to each other. So what is going to happen. There would be a requirement of a person who should carry a message from actor that yeah actor has completed the certain part and now it's your part to do the thing. And that is what a broadcast message is doing, basically taking a message from a sprite and coming to a different sprite and checking and telling you this is the message you have to do. Okay, so if you I have received this message. Now you have to do a certain task. So I hope I have tried my level best to explain the thing but still, if it is not understood not a problem. I'm always very happy to come up with new ways to explain the thing. So we can of course discuss the things you can share your doubts and telegram group. But before we go ahead I would request Tina to please come up please ask some of the questions which was shared by the educators on the last year. Okay, so the first question is, will we get separate certificate for both the level. Yeah, of course there would be separate certificate would be shared which would which you would be able to download it from your own dashboard. Once you have completed the activity. So, once you have completed all the activities within 24 hours your certificate will be generated inside the dashboard. And both I should say the step there will be separate certificate for both the bootcamp level. Okay, the second one, can we add schools to the hand control game. Yeah, of course you can add and have scores to hand control game and for that basically you would require a concept called a variable. So what you can do is I'll just try to showcase a small way to do that. Just a sec. Okay, so what you can do is you can create a variable called as a score. Okay, and when you have added a score. So here you have will get a score. So this is a variable. Now you want it to be edited. Okay, so we have some blocks. So this is going to get it edited. So of course at the initial level we would require to set it as zero, but here you would require some conditions, like if certain thing happens. Okay, then you have to change the variables by one. Okay, so now this is the condition upon you that what you make it, like if I would say, just assume that I want score to be increased. Okay, so my Toby is touching this right arrow key. And if I make it in reference of the last activity. So if in case my bullet touches the monster, I want my variable to be increased. Okay, so what can be done here is of course you can use sensing. Okay, and check whether the Toby is touching the. Okay, so instead of mouse pointer right now it's a button for the here you can just check the user sensing block and check if your hobby is touching this, then the variable should be increased by one. So right now it is not changing, but as soon as it gets touched. Okay, let me just touch it. Okay, so basically instead of variable we have to change the name that is a score. Okay, so right now we have a score. So, if it is touching, of course, the score would be one if it is not touching, it is not going to increase. Again, I'll just show if it is going to touch it, it is going to change it one. So right now, if you see it is not working on forever, but if I put it around with forever, so it is going to work automatically and now should see how exactly with high speed, this score is going to change. Okay, let me just stop this. Okay, it is not touching. Okay, it's true. And so basically, you have to basically make a condition. Okay, and then you have to change the score by checking that condition using change score by one variable. Okay, of course, it can be used in forever. But of course, with that reference, it would be a little different. Okay, so you can try. And if you in case you find any trouble, of course, you can share it on telegram. Yeah. And we have one final question, Ayush. Why we are using Victor blocks, why can't we use scratch. Okay, so that's actually one of the best questions I like to answer it's very, very precisely. Of course, scratch is also a block base. But in Victor blocks you are getting so many other extensions, as well as you are also able to impact with the robotics part. So here if I go in extensions, but these many extensions are first you are not going to get in any of other platform that is face detection, object detection, human body detection, machine learning and many more NLP and physics. Alexa, if triple T and many more extensions, apart from it, as we have a wonderful hardware with us that is the different hardware. Let me just select a board. So yeah, there are different hardware which can be connected with the crypto block very easily. The very important answer to this question is for Victor blocks, you can always come to us and ask any doubt, any problem, any suggestion, we are going to support you. Of course, and you can put it forward on support at the rate. The stampede.com. And I won't say why not to use scratch but I can let you know that why exactly the crypto blocks is far better than any other platform right now in available in for the schools and for the educated. Yeah. That's all I use for today. Okay, so yeah, as I was seeing an answer your picture blocks or sorry, class three has also have extensions but they don't have all the extensions because some and many extensions have been created by our team. Okay, and you would be able to use AI of and many other things very easily in crypto blocks and of course we'll be able to get our support for understanding or resolving any of your problems. So I hope that is pretty clear, and it is going to also very well work in your Android phone. So I hope this is very clear to everyone and I hope you have all enjoyed today's session, you might have learned the new thing. And looking forward to connect you tomorrow to learn about the machine learning that is going to be very very interesting. So thank you everyone for joining us taking out your valuable time for us. Do not forget to fill your attendance form survey form. After the session and submit your assignments as soon as possible. Thank you everyone. Let's connect tomorrow again at the same time. Thank you. That was a great session today. We all understand the basic concept of natural language processing and how exactly it's functioning. And hope all our educators really enjoyed today's session. We have a request all of you fill the survey form at the earliest and thank you all our dear educators for joining. See you tomorrow. Take care.