 So, what is summer internship? So, what do you think? It's just like a course, it has a course that you are doing. So, you have to ask questions, you get answers, same. So, have you understood, I mean, have you seen what is, I mean, you have tried asking questions when you were actually doing a course and then you keep on asking. So, then you understand. Towards the end of the course, you come to know what is programming, what is artificial intelligence. But before that, you have a very narrow or very vague idea of what that course is. It's very singular situation. If you are done internship somewhere else, then you might understand. So, can we ask what is what? So, what does the dictionary meaning say that? We can ask so much, we can ask this also. So, if you know this, then you know many other things. So, what is it? So, difficult to define. So, this talk is actually I spent five months or six months creating this talk. So, it includes algebra, it includes formal logic, it includes system dynamics and it includes something else, which I don't remember. So, it's like you want to define something, you need some rules to define that. I mean, I think you have gone through such definitions, wherein you want to prove something, you get some already written axioms for that. But if you want to prove axioms, then you need a higher language. So, then you have new axioms for that language and then the new axiom over here becomes a theorem in that language. So, that's how it is. So, it cannot be understood unless and until. So, the only way to understand is what is it? It's not what it is, right? So, it's not something else and you know what it is. So, particular object or anything. So, it could be a concept, it could be anything. So, the fundamentals of any language, you have six wives and one husband. So, what, why, which, who and these people is to fight all the time, you know. So, the what, why would always ask what question. Likewise, why question, which question. So, all of them would fight and then they would understand something. Anyway, there's no story like that. I try creating it myself. So, what do we finally understand, you know? It's not only asking question, just using what as an interrogative, but other things are also important. So, if you get stuck somewhere, just remember this six wives and one husband. So, you ask questions. So, question has to be, you cannot ask somebody to tell me what could be a question, but you have to generate that question, right? Then only the understanding is there. This is the title and I was scared to put this title on the front. It is universal paradox, circular oscillating self-referential. So, take an example, x square plus 1 equals 0, x equals minus 1 by x. So, if you want to put a value of x as 1 and then you want to know what is the value of x, you get minus 1. So, it's like a paradox. So, it's like what you say I, imaginary. So, it's circular. The never stable, it goes from minus 1 to 1. It keeps on oscillating. You will not find this concept. You will find it in one of the books, which is mentioned in the references. And he considered a Boolean algebra is actually not an algebra, in that, you know, where he tried to give a lot of supporting reasons, saying that Boolean algebra is actually not an algebra. And you have, I mean, this is just included for us to remember. So, if I say I is imaginary, but with respect to real values, right, then phi and not of phi is also imaginary with true knowledge. So, if I have knowledge phi, I can also say not of phi, not of phi, not having that knowledge. It can complete my knowledge, because it's like A versus NP. Does it mean that you know P, then you know NP, or do you feel that if you know NP, you know it is something is P, right. So, that's how we learn. We learn so slowly, we segregate. So, we are trying, you know, we are trying to solve, I mean, people are trying to solve P versus NP problem. Why they are trying to solve that problem? So, if you know something is P, then you know something else is NP, right. Actually, a mother clearing CS, which I mean, it cannot exist, actually. Phi and not of phi cannot exist at the same time. It's just a normal way of the declaration. So, paradox is important. In fact, but you know, when something is contradicting, we don't feel it good, but if you talk about OMS, load of contradiction and through those contradictory sentences, you know, OMS have become so widely world-known that people use them all the time, right. So, you can see this. Our greatest songs are those that tell of saddest thought. If can it happen, I come back to where I have never been. Can it happen, right. So, in that context, actually our mind feels very happy when there is some contradiction, I think, something like that. Are you happy when after reading this, right. So, what we have to do, we have to come out from this happiness, you know. If you have to be happy, then you will not be able to ask questions. You have to come out from that. So, we have not evolved programmatically or whatever. So, we have a lot of paradoxes around. Some people tell us something, some people tell us something. We take it, we change our thinking, we change our definition of a particular concept or keep on changing, you know. So, always evolving. It is not the same. I think five years back, probably what you had thought of programming is not the same now. So, it is changing, you know, even after ten years, twenty years it will change. So, but then you cannot change so much that programming does not become, I mean, should not become non-programming, right. It is hard to change paradoxes around. So, we have to understand the six W's and one elsewhere. So, how do we do this? I mean, it is not that easy. You know, you have to keep on asking questions. Sometimes we get bored because you are not used to asking questions, right. Sometimes you may find, so, you know, you may have to write something, but you are not used to writing anything. So, but then the simplest way to do it is write some sentences and then keep on asking questions. That is how it becomes bigger and bigger. So, right. So, it is nothing different than doing, you know, un-tech thesis or your PhD thesis because you have to keep on asking questions, right. You have to get answers for that. So, answering is not important. Questioning is important, actually. If you do not get an answer to any, to your question, it should not be said, you know. It should be much happier because you have a chance to find an answer to that question. You keep on asking that question. I mean, I am not saying asking question. Everybody asks to yourself, right. So, that will, there will be curiosity there. So, then you keep on searching, searching answer to that question. So, answering is not important. Otherwise, you are like a machine. Ask the question, give me an answer. Give me the answer. Give me the answer. The IQ is very high. IQ, I mean, very high IQ is also not important, right. You know that. So, the way of keeping says, I keep six or honest serving men. They taught me all I need. Their names are what and why and when and how and where and who. Okay. This is mine. So, you need six couples and manage along with their compliments. Not just the what, but not what or not why. Okay. Six couples. Not correct? Five, isn't it? Five of those. No. The other six there, five of them. No, down there or up. Down there is five. No, no, no. This slide you also mentioned five. One is lost. Six. Which is lost? Oh, you are talking about the Rudyard Kipling. Yes. Oh, I don't know why we not consider one of them. Which? So, which is lost here? Which is lost here? I asked you a question which is lost. I don't know. Which is lost? Question is the answer is the same one. Is it that a paradox? Your question is equal to the answer. So, when paradoxes happen people like it. So happy. Okay. So, this is a, I try to put it as a one day cycle, but it could be two days, three days also. So, this is the knowledge that you are trying to gain. So, everywhere there is, if you work hard you understand and if you understand like it and then you work hard again. People say that some of them they say they, I don't like mathematics. It's because they don't understand mathematics. Because they have not worked hard, right. So, if you work hard you will understand. So, it's like that. The cycle play also almost same. What if you, if you keep on working hard? Very hard, very hard. What happens to this? Your cycle is somehow this should get completed. It has to be balanced, right. Otherwise that will be too large and so you also learn something from here, from here, from here. Right? You have to enjoy. So, these are arrows are important here. You can see the arrows. They are not the quadratic thing. In fact, if you keep on enjoying more, then this cycle, this will not run. And because this will not run, you are, if you keep on working, suppose if you keep on working very hard, if you keep on working very hard, there is an arrow here of the arrow. This has to revolve, right. So, arrows are important. This is the typical system dynamics way of representing. So, how to think critically? This is from Polia. Anybody knows Polia? Polia? He was a great Russian mathematician. Great. Why great? Because he had many theorems. Can you name one? Sir, I don't know exactly. He has written a great book. You have read at least? No, I just read the name. How can you assume that he is great? Anyway, he has given something to us and this is, I think 1930s were quite very old and it is still valid. Not only him, I have seen many, many people do consider this as a chart. You can put it in your bedroom, wherever you want, you know, big screen. But it's not that you have to follow everything. You can read it, you know. So, first thing is you have to understand the problem. There are some steps which are given there, right. But if you read it and if you relate it to what you are trying to solve, you will feel that you are very close to it. But if you miss out something, if you miss out something and just go and read this, you may find that in that. So, just the reference, okay, so that needs to be, you can put it up somewhere, read something. Like a prayer, maybe for some people, not for all. So, you can put it unknown. So, there are, in a problem you have three things. One is unknown, the data and the condition. I would put out this slide somewhere else so that you can read it up later, you know. So, the first is how to understand the problem. Second thing is how to find some sort of connection between the data and unknown. So, suppose I am trying to solve a problem and if you feel that data is not sufficient, then I will try to reduce my data and then make it, you know, or I try to change my unknown in some form wherein I try to put it somewhere, okay. So, some sort of, you know, a little bit high level of, you know, I need to change something somewhere, I need to change. So, you are trying to evolve, you know, all the time and that is why you ask these questions. So, like, you know, you can ask an auxiliary problem, you know, if you want to, I want to solve P versus NP, it is not possible. So, you better solve something else, no. And why you waste 100 years solving only P versus NP? Just solve something else, there are auxiliary problems available. So, you slowly have some transition in different different domains slowly, right. And of course, you should need a plan towards, and then you, you know, it is like, and then you need, you know, it is like, it is like a cycle actually, it is not from top to down, like one, one go, but you have to go back actually, you have to see whether it has really what you have reached somewhere or what any solution that you have acquired and is it really fitting back. So, you have to go back, sometimes you have to iterate in the second, in the second step also, keep on iterating, it is like that, ok. So, these are my views, maybe you can take it or you do not want to take it, you can take some steps, purpose and motivation, you need a purpose, you know, when you want to do something and you just do not do anything, keep on doing just because you feel that people are working, I also want to work. So, there must be a purpose, there are some motivations, some good motivating examples, where you get fascinated, like for example, my friend he has said, he has heard about Polia, but somewhere he has started, you know, like so, if we understand it a bit then consider him as a great mathematician, yeah. Review of the existing literature, so says with proper criticism, of course, because you cannot, you should never believe in anybody, never believe, even if the author says that it is like this, no, I will have to prove it myself, then only I will figure out what is right and what is wrong, so what, it has to fit inside first, otherwise it will be very unhappy person, it will be very difficult to identify, then only the problem is, problem is very crisp problem, not just something called as whatever, big data analytics, it is not a problem, it is a big field, it is not a problem, problem has to be very crisp and solvable actually, putting forward specific and non-missive questions, any of those hypothesis, it is any of those hypothesis, in fact, there is something on my website, so when you can go and read, I have drawn a diagram, you may like it, if you don't like it, don't even look at it, so almost the same, tackling the problem with critical analysis, giving various solutions to the problem I identified, not just one solution, at least 2, 3 if you can give and then you choose, which one case study of course required, you just don't say that this system is going to work, but if there is no case study done, then you don't know whether your system is going to work. So, then you have to demonstrate to somebody, proper documentation is also important, also future, because if you are thinking that you have built up your knowledge on somebody else's shoulders, like reading from somewhere, then it is what future possibilities, so it is your duty to write there, that this is what future possibilities are possible in this area, so it is your duty and it is more than solving a problem, then a clear cut presentation, so that even a layman can understand, so I will ask, no hi-fi terms, analytics, data, they are very difficult to understand. Data sources, journals, non-graphs, renowned books, white papers, technical circumstances, software documentation, even newspapers, you may find something interesting in newspaper and you want to solve it and then it is applicable for everybody including me to do an understanding, that will be it. So, these are my references, in fact the circularity definition is very hard of computer programming, anybody who knows can you tell me now what contribution he has done. He has wrote a great book on algorithms, it has 6 to 7, 3 to 4 volumes out of computer program and he has known as the father of algorithm design he has invented many which algorithms Sir, actually he has means he had made the bigot notations and all that famous for means, analyzing the algorithm. So, you have read that book? Sir, I no, I have not read it. No. I would not get hands on that book, actually I have. Anybody read that book? It is difficult, it was a Bible before when no books were ever done, but he defines what is called a circularity do not surprise how he does it. In the index he defines in his book he writes circularity. So, somebody has to go in the index and find out where is it defined, right. So, there he defines C circularity Sir, talk here Then down somewhere C D C circular definition in the index how beautifully he has put it right. Lower it is mentioned, but in the index it is there. C D C circular definition With that I think we should understand what is circular definition. We will need to study a proper definition to understand certain definitions, right. So, we saw internship projects, but I am not the only person there are other people also. Sir, the big thing I mean the broader perspective it is more about cloud and cloud is also a jug and you should ask me what is cloud? Big data, it is more about data which is structured as well as unstructured or any type of data that exists in the system and what you want to do with that data. There has to be some way to analyze it in some way that you can find some meaningful information from that data. So, that is the whole process. So, there are lots of you know lot of efforts going in different all over the world and people are trying to design some standards. They are trying to design some move DPE or something like that. So, they are trying to say that for this particular platform is a edX platform open edX platform partly we will modify to suit the Indian requirements and so you know efforts are there they want to standardize on say even the schema there is no schema first of all. So, you would like to involve some schema. So, there are so you may consider it or we also have involved some schema. Our schema is different than the mighty schema. So, you can you are free to consider one of them. So, because you need to know which one is the best which one is better not best. So, there are logs which have you know when some students actually navigate on this platform. Logs get generated every event log. So, what do you do with those events? Now, you have to find some way out you have to you need to know something about the student or you need to know how the course is running. There are a lot of pictures you know the teacher would like to know you know the course or for a particular student or a group of students who are they performing. I mean so of you know analysis or analytics or analysis above possible you know you can have a very basic analysis first and then you can build on build on abstract them abstract them all right. So, there is a inside platform I think of in fact a partner covered a section of this in your first site. That eventually it is related to the ID Bombay X because outside the scope right. No, it needs to be given eventually that is why I am asking. Eventually means before June or after after they leave or before they leave? No, I think they can provide some proof of concept for eventually after June. My only concern is I am adding ID Bombay X. I haven't received any request from you. ID Bombay X? No, no, no, no. I will be talking on the platform. I mean just if they are able to show me on open at X code that is fine. I have not any impact on that which is not gone by me. Right. Industry speaking. You have data in the same. Then I will be interested in what you are doing in coming inside. I will never touch your system. I will not touch your system somewhere. Somewhere else. Somewhere else. The rest of the IOS is coming in. That is certain. So there are some basic some information. For example, how many so how many males or how many females. Such type of demographic information or how many students have answered this quiz correctly or that quiz correctly whatever it is. Very basic information there. But there are some interesting questions which we can ask like how many students and how we are learning resource A. So it is also possible to ask. If you want to know how many students have actually viewed that resource it is possible to say the opposite way. So if a student do well on activity B, so they do well on activity C, is it so then you can see if this is true then you say something right and if this is true then you can say something. So we do not know what is that. You can think what is that. If they solve excise the do they also solve excise E, they may be related you know interrelated. So maybe quite very similar excises. So if they solve this they are able to solve that. So these are some questions which you can put forward. There are some 30-40 questions written down also. We will provide you that. See which courses have you learned a lot of learning resources and which of the questions please are not answered correctly at all. So they can tell the teacher in some way that this quiz was very bad so do I need to put up another quiz or maybe the particular concept is not understood properly or set of concepts or there is a dependency between this and this. So if they are not understood to this then they are also not understood this. So it is a way you can identify some relationship between you know entities and then come with an outcome. Then the teacher has to decide what to do with it. So the teacher says you have to give a new quiz. My system says this is my information and my system says you should give a new quiz but that is not correct. So the teacher has to analyse too. The system will tell you some analysis here and there. So these are some other questions. So you can think. The new quiz is Dr. Fartan's special requirement because if he finds the quiz people have failed he dilutes it and gives a new quiz so that people will pass. Yes, another problem that is called make up. Make up, he always has got make up. So everybody makes up towards that. You can also find the time required to solve the particular problem. So many questions should come in your mind. The first day of your internship, I don't know not today, if I find that particular group, that group has to write at least 200 questions. Is it okay? So this is an open space is one thing that we are trying because the whole cloud today that we have, cloud is something the bunch of VMs virtual machines or there is a platform where particular environment is there and in that environment you have bunch of VMs, virtual machines instead of physical machine I can say 10 number of virtual machines in my system. I can have 10 VMs inside this. In this system I can have 10 VMs I should be able to configure it instead of running it on this machine or instead of running on the whole machine I want to run, say for example this software on the one VM So I should be able to configure all of those requirements. I need as much memory for this process. I need as much memory for hard disk space or as much storage should be attached to this VM It is all everything virtual So the cloud open stack is one product which allows you to do that Right? So I made a mistake. I did everything in the morning He says the description is the same of the previous project that we did Oh! It happens sometimes Project is different Project is different The name is there The number is the same The number of intents is 8 That is correct I should be correct somewhere No, no, it should be 4 I think 4, 4, 4 So how the net it is correct 4 is 4 And then I have everything somewhere here Whatever is there on the net is there Whatever is there on the net On this time Just to show you There is also an open stack A lot of other peripheral elements where you can have your separate storage and you can configure it as you want So it is a very large product We have not used it to that extent So we would like present cloud that we have is actually put on VMware in the open stack So we would like to have the whole platform using open stack as a proof of concept we would have to demonstrate either EDX or any of those because we already have specification or platform deployment So those you can consider So some machines would be given to you of course As you can see Swift is one where we are interested because Open edX platform actually connects to S3 services And if you don't want you can always stop the services But what happens is that you don't want the data to exist You would like the data to exist on storage clusters or separate clusters where you can manage separately rather than putting everything at one place So that thing becomes highly scalable because you are not loading a particular machine The cloud will take care of all those So performance is one thing which is sometimes very important because you may feel that you have a product today which is running you may feel that it will support even 5 million students or 10 million students but then it may not scale to that level So what do you do? So you have to actually test out something Is it possible that do we need little better hardware or do we need more clusters So what is the solution? So the solution we need to find out You cannot say that my system will scale to 1 million 2 million just like that There must be some, there should be some test results and saying that I have simulated it in some way or the other At least an experiment is performed and I feel that it can hold to this this much level or whatever So we use the jmeter here the word is also being used You may also need to consider work load behavior These are just an extension to what is, when you talk about performance issues what do you want to really see Is there a uniform access or skew access to certain hardware and certain disks which are there to find some problem first and then you want to solve that So you need to probe into something I am just giving you some very broad overview So is there some variations over time Is it a cluster Suddenly is there an unpredictable load somewhere Is there a problem So where may be the afternoon at 12 to 2 watching movies Do you know when is the peak load So you need to find certain large burst interval load So you may also find what type of workload is it Is it just a video access Or is it just from the data actually you can figure out how much video has been used how much slides have been used or how much problems or quizzes So you will have a fair distribution once you know your data data is there so you can extract this information So there are some benchmarks already available So I put some references there And they have actually tried to use it So this is a VCH type of benchmarks which is generally used for benchmarking So OATB systems No this is a decision support system they really use where in you would like to know how much of processing power can it have So this is my hardware capable of doing that So it has been built on already existing benchmarks and So and even if the reference is there is a book also available it can read a whole lot So if you are talking about Hadoop when you say you have So you use on that architecture in the beginning that big architecture So that architecture whatever if you say the platform the platform does not include what you say Hadoop and all those activities It's not there So the platform is separate And if you want to do analytics and the basic analytics which is there NEDX does not use Hadoop at all There is some elastic search which is there and you know bunch of which I don't know but they don't use Hadoop So if you put Hadoop for this can it do real time or near real time you know analytics and use some information to the teacher like you are surprised suddenly right when you are on Gmail suddenly pops up and tells you that do you like this but how does it know it tries to find some whatever you have searched before I have a question Why real time information required in the MOOC scenario It may not be required It is not required It is not required I don't know Human instance The whole purpose of MOOC is there are so many people all over the world they have been even one week whatever they want Suppose I find that people are answering questions now right now and there is a deadline which stops in another one second or one minute and if I know that I they are answering I would like to extend the time it is possible I would like to extend another one hour for them if they are answering now I would like to do that I don't understand you have given them one week they want to do in the last seven minutes so it happens that people answer towards the end only now the time saying that one hour is there they will do something but if I find they are not able to solve can I estimate how much time they will take no can I estimate there are issues there I have listed where I am now tell me you extended isn't it unfair to people who have answered how would it not unfair I was not sure I did it because the time limit was there now we are extending it it is unfair it is very simple you can't change the rules okay so you get changed based on that let us we will use that system you change the time and then you don't consider them as marks you show them that I have changed but you change it so that you get some information about them that they have taken this much amount to solve this at least they are understanding you will have to know I don't think real time is required so real time is you will not need because unless until the stock market something happening I mean you need otherwise you don't need so even if you have a requirement which is completely different I answer a question in a physical class depending on the answer I should do my explanation there is a real time requirement I agree but not here don't extend don't extend the requirements of the earlier system to a new system it is completely new nothing to do with classroom requirement doesn't going to move two different ways of education so this is our fourth project about replicating I don't know in fact works are part of load days and I just put it up here so did anybody understand this can we have a discussion and I am ready to answer any of the questions no I know what it is this is the this is the this is the same with this blended model oh that is in relation to blended model so I am not not mad so somewhere the relationship is something else the blended model earlier add a lot of servers and the data is missing and the reason why is putting it up is when we did the implementation I told you very flatly that I will not do real time I will do in a batch because I have come from that fundamental thing that real time is not required anymore so whatever I had implemented tried to implement or to implement it completely was a batch yes that's why he wants a real time system implemented for a distributed data base which he can use in the future so there is a main system and there is other systems which are connected the main type was the concept which we are trying to implement was the whole data is put in by an IIT professor some professor who is delivering the book the same copies there in every college same copy of the course the college faculty can add to it his own lectures his own quizzes etc that's why it is a copy now college A whatever is added is college A proprietary I cannot share it with college B nor can I share to this thing so the concept was there are different sets of data available everywhere but the central faculty has to know what is happening in the local colleges the data has to be seen it will pair kind of thing it's not full network because I had implemented it here I said no so something happens on one side you have the instructor maintaining the whole the platform for that course those mini lectures whatever he puts up it has to be seen somewhere and then some discussions happening from the other side maybe you would like to see that was his requirement I said no I hold it this college is charging money for training its students you may have your own water idealistic system but the world works on money this college will not agree that water money it has spent or effort it has spent should be available to its competitors academic the business is there all colleges run on business so the system has to be designed like that but there are people like DBP who does not like business he has a strong know-how to business but he is dealing with businessmen who is selling education he lives in businessmen to get funding he wants to absorb this I have been trying to tell him that this particular model he has to charge the colleges he does not listen I cannot help him I am against all colleges he charged 2 points delivered nothing you made them too much and they do not give anything so the fixed project between Moodle and edX edX edX open edX open edX do it from id compics you should do it id compics after I have delivered the software to the system administration to run after 100 times the production system maintenance is outside my group you can add features to it independently I have no problem do not come to me to add I will give you I will give you I will give you translation these are my two four translation management is included both are related of course that is what she said two activities we are working on translatability of all course content automatic translation automatic translation if you want I will give you automated here I have done it in a project I have no problem that is Google and for this particular purpose we are doing it officially because Google for educational purposes Google gives an API free for my purposes I achieved I thank them they have translated Google and I thank them officially for making their free source available actually in my earlier time I changed I said I have thank you officially for making it available without adding a capture if you want to protect it add a capture everybody knows that if you don't add a capture I will change and Google can't claim they don't know capture technology it's obvious they want me to change to don't leave the door open don't leave the door open you know there is no capture where is that R? somebody found an R somewhere no R is going to come later on R is being called but that is only one project and she will not get too many people okay so R is just a statistical package no I hate you are not doing the R now you are doing QQs stop the R person because of public demand is going to come and talk about R so who is that R? it was an R oh now I got it you don't have to talk I am faster to come here fine so I finished my talk here I feel one you can go through my slides later what it takes will be provided so I will be putting up it here last year we used this for all project description you are the key no no that is brainstorm something you are done that was better we did not have time we finished the project descriptions only very late I told him we were busy with lot of other things okay so that is the link so in fact it has been under construction last year at least I had 400 to find a reference and this time I have less I am coming with your heart