 The purpose of this class is two-fold, one is to make you aware of the group discussion approach and also to use that approach to get inputs from all of us towards identifying sub-parts of the larger problem and seek possible solutions for that problem. How many of you have participated in any group discussion? Many of you. As a part of what activity? Placement? All of you are familiar with group discussion as a method of placement. Some of you who or some of your friends who would have appeared for interviews for admission to the management schools, they would have faced group discussion. It is also a kind of placement, a different kind of placement to being offered as studentship in that college. Unfortunately, the main purpose of group discussion is neither. The group discussion that is conducted for placement and the group discussion that is conducted for admission to an MBA program are merely assessments to find out whether you are good at group discussions or not. Now, here is a dilemma. If the only time you have participated in the group discussion is for an assessment for either employment or for admission, that means you have not really understood the objective and purpose of the group discussion. So, any idea what is the objective of a group discussion outside the placement and admission? Come on, many of you have participated in group discussion. This is an exchange of ideas. Exchange of ideas could also happen between two individuals or even if there is a crowd of five or six people, a general conversation could also lead to exchange of ideas. You are right, but a group discussion does something more. A group discussion is usually structured. So, what do we mean by structured? So, the group discussion is structured. What exactly do we mean by structured? So, how do we structure a group discussion? Come on, you can cite from the experience that you had in the group discussion. Who starts it? Anyone from the group? Is there no moderator present during your group discussion? It depends. So, when a group discussion happens within an organization outside the ambit of employment or admission, generally some kind of either a hierarchy or a pre-appointment prevails. So, either the senior most person initiates that discussion or a person pre-appointed to moderate that discussion controls the discussion. So, usually one part of structuring is that there is a moderator. In fact, the advantage of moderator is to fold. The moderator is a person who permits everybody to express the views but does not let any single person hog the entire time and avoids repetition of points. A moderator is also a timekeeper and a notekeeper. So, it is the responsibility of the moderator to write minutes of that group discussion at the end. And the moderator must say that, okay, we have exactly 50 minutes for this group discussion. We must preserve the last 7 minutes for conclusions and therefore the time shall be allocated like this. Why is that important? Because otherwise each one of us who is so eager to say something, you know, all of us like love to listen to our own voice, right? It is a natural for all human beings. So, we would like to speak continuously for all the time but in the process we might curtail the articulation of some important points which somebody else might have thought. That is the reason why the group discussion has to be moderated and that is the reason why the moderator has to take care of these things. Plus, at the end of the day, within 2 days, all of us may tend to forget some or all the points that were discussed and any conclusions derived and therefore some notes have to be taken on what. All right. So, today we shall have a brief group discussion. Not for the purpose of employment, not for the purpose of admission, not for the purpose of spending time in a class, but for the purpose of understanding how a group discussion is an extremely important thing for multiple people to make contributions to certain thought process. In fact, the topic I have chosen for this thought process is precisely accentuating this very activity of group discussion, namely building and nurturing collaborating communities. What do you understand by collaborating communities? Community is a group of people. Communities is multiple groups of people. Multiple groups of people, each group individually collaborating with all members of that group for achieving certain purpose. So, collaborating group is a group which collaborates in our context for solving a problem and when we say multiple such groups, we call them collaborative communities. The best example of collaborative communities are open source communities. How many of you have gone to jithab or sourceforce.com in your life? Many. Now, whenever you look at a project apart from finding out what that particular software piece does, you might also out of curiosity have seen the people behind that project. One of the hallmarks of those open source projects which succeed versus the others which do not is the establishment of large collaborating communities behind a particular project. While Linux competes with the best of the operating systems in the world because Linux has an extraordinarily large community supporting. You take other well-known pieces of software that we use which came from open source, take Moodle for example, take Drupal for example. Now, these have large collaborative communities. There are international conferences where people who are all volunteers actually pay money from their own pockets to attend such conferences, exchange views and collaborate further. So, these are the examples of collaborative communities. I will tell you another reason why this particular topic is of importance to us at IIT Bombay and for some activities that we are doing for the whole nation. I had briefly mentioned I think when I was telling you that those of you do not have a seminar registered for might want to work on literature survey on some other topics that I will give. One of the objectives of today's session is also to evolve such topics and in the process everybody to understand what exactly we mean by this. So, let me tell you where we need collaborative community. You are familiar with the fact that we are undertaking large scale teachers training program. We trained 10,000 teachers that time, we have trained 1 lakh teachers. You are also aware of the fact that massive open online courses have emerged in the world and IIT Bombay specifically is offering courses the small difference. We want to create these courses in multiple Indian languages starting with engineering education, going to college education and then going to school education. That is the objective. And in the process there is a commitment that IIT Bombay has established that all the knowledge content so created in the process of offering MOOCs will always be released under creative commons in open source for every human being to benefit. Now, if you want to create quality content, can it be only a job of a few teachers who design that course? There are groups of few teachers designing such courses in every university. IIT Bombay is no exception. Of course our courses are better because our teachers have perhaps more experience and are better players to understand whatever is of significance in a course. So probably they give better courses. But does it mean that they are the only ones who are knowledgeable in that topic? No. Does it mean that for a particular problem the best explanation can only be given by one of the IIT professors? No sir. There are many teachers across the country, many students across the country who can write better explanations for a particular problem. Better examples can be created by them, better quiz questions can be created by them. Where is the opportunity for such large number of possible contributors to contribute to creation of such knowledge and its incorporation in a course? The point is, thousands of people might be writing examples and quiz problems and so on. Thousands of people might be writing explanations and examples but they do not come into a textbook. Textbook is written by one author or two authors. Exactly the same thing about a course. When massive open online courses happen, there is a possibility that a single course will become so predominantly useful because it is meaningful and useful that it might prevent innovation from happening at other places among other students and among teachers. How do you prevent that? You prevent that by saying, look this is a course, we take this as a beginning course. It is a course on thermodynamics. Three great people from Mechanical Engineering, Gaitondek, Bhannarkar and Milindathri do that course. Their course is well-known across the world by the way already. Yet if they say, now look here, this is the course. But these are 10,000 people who have taken the course, teachers, students alike. Make contributions of your own. Now people will make contributions. So this is the glimpse of a collaborative community that is being created. Unfortunately this act alone is not adequate because many of those contributions could be done. I am an enthusiastic person from a small village like community. I am good at technology but my way of explaining is so rubbish. I cannot write correct English. So what I write is useless. There are other person on the other hand who explains better even than the teachers at IIT Bomb. How do you distinguish between the two? Well, somebody should edit. Somebody should review. Agreed? This is what happens in technical literature that is published. There are reviewers, blind reviewers, etc., etc. And then a paper gets accepted. Here we are talking about a large number of small content being submitted by individuals voluntarily and the need to edit them. Can we request Gaitondek, Bhannarkar and Milindathri to edit such submissions? Of course we can. And they will do it as long as such submissions are about 5 or 6 a month. But if there are 10,000 students and if all take this challenge enthusiastically, there will be 10,000 submissions. My colleagues may resign from IIT Bombay if they are forced to look at this. They cannot. In fact no human being can. In the automated process of assessing the meaningfulness of the submission, so far not. You are all computer science students. Hopefully someday you will create that. But today we don't have. You need human beings. So what is the mechanism to do that? Have you heard of crowd sourcing? And have you heard of peer assessment? In a way, when I requested you that you should record your 5 minutes of public speech and also comment on some other colleagues' public speech. When you comment on some others' public speech, you are actually doing a peer review. Peers means equals. People who are doing something similar. So while 10,000 people submit their assessment, we randomly pick out these assessments and randomly pick out people and say, look, you 10 people will assess these 5 submissions. Each one assesses 5 submissions. Gives a grade. ABC, etc., whatever, whatever. Now look at the objective when you are operating at large scale. You are not interested in the ranking all 10,000 submissions. First, second, 4,458 on any parameter. You are not interested. But you are interested in gross filtering of these 10,000 submissions. Not very good at all. Not to be seen further. Reasonable and the best. Because you know that the best have a chance to be finally approved for inclusion in the main course. So you are looking at the best assessments, the best submissions and the people who make those best submissions. Because they are capable of making further submissions like that. Agree? The point is, without any human intervention, without any human management, how do you initiate this activity? How do you permit peer assessment to happen? How do you automatically collect all the results? How do you make groups of people who submit the best okay and useless kind of submissions? And more important, how do you then start assessing the assessors? You have made groups of 10 people saying each one assesses 5 and so on and so forth. Now imagine that out of all these assessment, which is a peer assessment done automatically, certain best assess things come up. What if I give the same set of best so-called things which have emerged for a re-peer assessment among the larger number of people? And now I am assessing the peers in their assessing capability. I can do two things. I can ask them to do a more refined assessment. I know they are all good or whatever. So I can say the best or this parameter, that parameter and find out their judgment. One, I can get a further filtered out components of these artifacts which have been submitted by people. And more important, I will know who are the people who assess meaningfully and correctly. Agreed? Now suppose I have an automated process which makes a group of such editors and moderators, prospective reviewers at the higher level. And of course another list of names which is a group of people who make such submissions. I mentioned 10,000 people which is not uncommon for a MOOC course to have participants. All 10,000 may not submit but 10% will. And of course if you make an assignment of one mark in that course which is dependent on you are making some submission then all 10,000 people will make submissions. Common sense right? So you have these submissions. Now imagine that you have multiple subjects that are being offered and you want to create such artifacts for multiple subjects. Best artifacts. Can you not apply the same mechanism and create multiple communities which are subject-wise organized? Now when you organize things subject-wise it might appear that you are organizing them automatically but hierarchically. Mechanical engineering, thermodynamics, fluid mechanics, whatever. Computer science, databases, algorithms, data structures. Is that always true? There might be a person who has extraordinary capability of contributing to a thermodynamic course and also a numerical computation course. So what do you need additionally? You need a mechanism for a taxonomy to emerge. Because you would like to classify. You will have to tag every artifact. Remember not thousands now, lakhs of artifacts have been submitted. Some of these are video clips, some of these are explanations, some of these are problems, some of these something else. They all pertain to different topics within thermodynamics or computer programming or whatever. Would you not like each such submission to be tagged by multiple keywords so that later on anybody can collect all the material which pertains to or which has linked to a particular keyword. So you require to evolve a taxonomy. Now we are talking about something really large scale. We are talking about kick-starting a process which will build collaborative communities in order to a, submit individual contributions. b, peer assess these contributions. c, come up with a ranking or a marking saying these submitters generally submit good quality thing. So you have a good quality submissions, good quality submitters and good quality reviewers. Good job so far, useful. But now you want to nurture these communities. Suppose we kick-started from IIT Bombay and such 8000 different communities get set up across the country, each community are 10 or 15000 people. And I am talking about a situation where we are not limiting ourselves to engineering education. High school maths, high school science, history, geography, accounting. Can you imagine the prospective? Now how do you nurture them? There cannot be a human control but there has to be some sort of a control. How do you do that? So you have to build leadership. In the world leadership is either self-imposed by an extraordinarily capable person who can ramrod everybody else and say I am the leader and then people meekly listen to it. Or the person is democratically elected. How can you automate such a process? Again fall back and learn about the open source communities something which you have never bothered to learn. How do these communities thrive? How do they continue their existence? They continue because leadership emerges from out of the same group. Somebody starts taking initiative. Somebody says I will coordinate this. Somebody says I will manage this side. Several other people find or like such a coordination and they agree to work in collaboration with that person. So that person becomes first coming big one. Leadership therefore has to evolve. And our system whatever we build must permit evolution of such leadership. So you see we have already identified a huge lot of software components that are required to build such a system. Such a common system does not exist anywhere. Although there are components of such collaboration. Gellol open source I mentioned. How many of you are familiar with Wikipedia process? All of you read Wikipedia articles. But you are computer students. I am amazed that you are not curious to find out how does Wikipedia itself run. Do find that out both in terms of the software that Wikipedia incorporates because it is a scalable software. Remember whatever we are doing has to be scalable. There are likely to be 50 lakh to 1 crore users out of 120 crore citizens of this country. That is the ambition. 50 lakh to 1 crore users spanning considering only the knowledge handling component that we are interested in the institution. Primary school, middle school, high school, junior college, college electives, research everything. The only requirement every content that is contributed is committed to be given out in open source and majority of the people work voluntary. But voluntary work in terms of contributions, in terms of peer assessment may actually work. Will it work for leadership level where leaders have to spend more time than normal. So can we think of some incentives. Incentives need not only be in the form of money to be paid. It could also be in the form of recognition, public recognition. So let's say the Ministry of Human Resource Development or AICT or some such body recognizes such contributors and such reviewers and actually mentions them in a list, sends that list in recognition, does anything else. Again this process also has to be automated although it will be more structured. So you need database of people, database of artifacts connected through taxonomy or database of taxonomy itself and the entire transactional process of submission, peer review, peer assessment, grading, voting, like, dislike, ranking. Can you not see practically all aspects of computer science dealing with data management, information management on large scale are at play here. There is a second component of building of such communities and building of such what should I say, open source content creation or OER creation. And that is the huge amount of research potential that the data itself will permit you to use. All these OERs or open educational resources will eventually find their place through peer review and editing, etc., into the course content. How nice it would be if five years later the IIT Bombay X course on thermodynamics still has Gayathonde, Bandarkar and Prof. Atre as the teachers or a database course done by Sudarshan and Sarada, they have still the names as teacher or a data structure course which Ajit Divan Ganeshan IR designed. But along with these names along with their material with sincere acknowledgement you have 200 other artifacts which have filtered out through this process and which are acknowledged and incorporated in the course. Wouldn't IIT Bombay be proud? Wouldn't it be very useful to large number of people? Now that is the objective. We are actually embarking on building such software. So Prof. Mausam, I don't know how many of you have heard of him, he is a researcher, he is a faculty member at IIT Delhi. We wanted him to come here but he joined IIT Delhi which is okay and Prof. Ganesh and I are going to build this. Another thing which I will quickly tell you, another large project which we are likely to actually venture into, this started by Ministry of Culture and they want to build a national virtual library of India. Basically all cultural heritage. The last 4,000 years of recorded history and hundreds of thousands of years of unrecorded history. So you take Bhagakaya or whatever which were constructed, maybe 50,000 years ago. Now you are aware of all the archaeological sites, you are aware of museums, you are aware of manuscripts, so many things, the entire cultural heritage of the country they want to capture in a virtual library and make it accessible to people. Can you not find a huge similarity between doing that task and building open educational contents? You have exactly the same kind of thing, you have short video clips, you have descriptions, you have hundred and documents which are digitized, you have photographs, and you have a taxonomy saying what is what and you need a community, for example, if somebody from the area of Hampi says that look I went to your site and this Vijayanagara Empire that you write, this is okay but my great great grandfathers had preserved a piece of paper which throws some additional light on this particular aspect of history in this year. Now that's a contribution. There has to be a mechanism to review it, assess it, authenticate it additionally and incorporate it. So you see once you build this kind of system, it could be used for variety of purpose. We don't know our history well by the way because it is not written by us. Only the other day in that abhyudaya I gave a talk on digital financial inclusion and I showed a photograph saying to recognize him and they didn't. Then I said he is Mohammed bin Tughlak and many people had blank faces. Then I said don't you know that he was one of the early kings in Delhi. Some people had heard of him. What special things he has done? Nobody knew. Two important things. He tried to shift his capital to a central place in his empire, Devgi. The second, he said gold and silver coins are horrendously costly so I could use copper coins, any minted copper coins. He failed in both, not because he was wrong but because he was ahead of his time and of course the mechanism he used because he was the king, he ramrodded everything, he ordered. You cannot build a trust which is required in a currency by ordering people to trust. But today all of us use Gandhi Baba's notes, right? When it says 500 rupees, we believe it is 500 rupees. What was wrong with Mohammed bin Tughlak? He at least was giving a physical copper coin with some weight but no, that time he could not succeed. Anyway I digress but what I want to tell you is that this national virtual library would permit the whole of the country with humongous diversity we actually learn and appreciate our own history and historical heritage proper. So I have only yesterday I took a decision that we will participate in this and do it, it's a big project but now again in order to make such project successful we need this. Now I will do the following. All of you, I will give you exactly 10 minutes. In these 10, you have heard all these dialogue. You now understand what is contemplated. Some of you might have been fascinated by one particular facet of this discussion. Some of you might have all of this in mind. I want you to jot down, I want you to jot down specific small problem statements which need to be solved as a component for building such a large system and I want you to mail these write-ups to me later but you will write them now because, विदगेशो बादगेशो after going back home you are not going to do any. So next 10 minutes use your pen and pencil use your mind reflect on whatever we have discussed and write down any, you can take any particular aspect of the holistic problem you can take taxonomy, you can take peer assessment you can take general crowd sourcing you can take voting, you can take content management how do you manage videos, how do you manage what should be the you can take content formats you can take the leadership evolution voting, you can take incentivizing you can take any one of these aspects or you can take multiple of these write down which are the problem statements such that each statement that you write resembles a topic for a seminar I hope you are convinced that this activity alone can give rise to at least 100 seminar topics each one worth its weight in gold because when contributed to it will actually result in further research and development and building these systems that is what we have to do any which way so you have actually a chance as kick starters to contribute to this activity and in fact if ever history is written I will make sure that names of all our students are written as the initiators of this activity provided of course you participate not just because I have a roll number belonging to this course in terms of what you write in next 10 minutes so write some of you are experts in hardware might want to contemplate on the sizing and architecture of the hardware that will need to be implemented obviously will require a cloud how many course in a cloud what kind of system that you would use data management system MySQL or MongoDB with distributed architecture or a combination of both you have to use a content management system such as Drupal as a base of the whole thing how many of you have seen Drupal or known about Drupal 1, 2, 3, 4 are you aware of Drupal 8 good Drupal 8 is a version which actually permits API based external development to be linked to the Drupal's inner content management system it opens up a Pandora's box where you can independently create applications outside which are not necessarily written using PHP so they could be for example Django and Python scripts running a transactional process outside or a web application outside which actually integrates with Drupal there is a conference on Drupal by the way in the coming week here in IIT those of you who are interested might participate in some of those sessions incidentally we have decided that we will use Drupal for the major content management but I digress so please apply your mind right now 1, 2 or 3 or whatever now you have 5 minutes left only I will keep quiet now otherwise you will get disturbed please stop writing by the way it was very obvious to me that 5 minutes or 7 minutes is too short a period for people to consolidate their thoughts and write them but the purpose was to ensure that your mind is initiated into that thinking okay now I would like the following to happen I would like you to spend atleast 15 to 20 minutes today sometime but today only and not late in the night before you forget you are all mtech or PhD students so that means you would be able to find 15 to 20 minutes time let us say before 6 pm today is that possible if that is possible I would like you to will set up a link on the moodle immediately I would like you to write a small snippet of your thoughts so they need not be exactly elaborate titles or whatever whatever some small snippet of one or two things that the ideas that come to you you all have to submit this and your submission will be taken as your attendance today that is important because tomorrow there will be a new a notice to all other people that if all other who are absent today have to pass this course they have to independently think on this subject by contacting you and having a group or individual discussion if required and ask for a cup of tea atleast if not for a treat if they do that and then they will have to submit a larger version of a similar thought process that is number one number two all of those who have not registered for a seminar topic or who have not already identified a topic for literature survey might choose from among the list which will be put up today evening by 7 o'clock and maybe and more elaborate submissions which will be done by others which Feroza will put up may be day after tomorrow you may have to choose a particular topic of your choice for the literature survey is that okay with you next week I am not here now I realize that you already spent about an hour extra in recording and here assessing your presentations you will be spending another hour and a half in the month of March doing the same thing that will have to be on a Saturday or Sunday so in anticipation of that one and a half hour and in recognition of the hour that you have spent we shall not have any sessions next week this will permit you to spend a little more time on preparation for your mid-sem which I think you should buttress my confidence in you by scoring one notch better in each of the subjects that you appear for your mid-sem but simultaneously you have to start work on the assignment the first assignment which is of collecting a large number of papers and perusing them with the quick reading as I mentioned as Sana Murti mentioned and preparing a list of 20, 30, 40, 50 papers which are relevant has to be completed three days after the mid-sem is that a fair requirement fine so in conclusion all submissions before 6 pm today be selection of a topic of your choice from amongst these by those who have not identified a topic three no lecture sessions next week because you will be spending time when you already spend time and four the first task of literature survey namely collecting 40, 50 papers conferences, journals, reports somebody might be doing a literature survey on some specific methodology or approach then you might write technical report citation whatever water you feel like but that preparation has to be submitted before three days after the end of the mid-sem is that clear we will just write this down and thank you so enjoy your mid-sem