The Era of Online Learning | Niema Moshiri | TEDxUCSD

You so back when I was a sophomore in undergrad I remember walking into my first genetics lecture I remember seeing this massive room filled with 350 students and I squeezed in found a seat in the room and I saw that there was the main projector screen which I was you see that's where the professor puts their slides but then on the side of the room there was another smaller screen and I was kind of confused what this other screen would be used for so the class starts the lights dim and then what I see is the lecture slides as expected but then on the smaller screen is the professor's face being projected to our class and I'm thinking that's kind of weird why would you project the professor when they're right in the front of the room so I look up and notice that there's actually no professor sitting there at the podium and it turns out that I was one of the first people in this linked classroom setting where they had a massive 700 person class had it split between two rooms and had the projector had the professor giving a lecture in one room and just projecting it to us in the other room and it did not feel very good it definitely felt more like I was reading some dystopian novel of some sort so clearly I was discontent with this as the lecture went on I gave a benefit the doubt the first lecture was just syllabus day so nothing too crazy and then one student in our room had a question to ask and then I was curious because the professors in the other room how are they going to ask this question they can't even see the student and it turns out that the mechanism was you had one TA in our room who was watching for students with their hands raised what text a designated TA in the other room that TA would distract the professor and tell him there's a question in the other room professor would pause a TA in this room would sprint with the microphone and effectively baton the microphone to the student so you can ask the question TA runs back to their seat class goes on as you can imagine this was pretty distracting it feels like you're watching the Olympic Games every single time someone's asking a question it was bizarre so clearly I was not very happy with this and I was wondering how exactly did we get here to this situation that I'm in a room with 700 other people taking the same class it should be fairly intuitive we have massively growing numbers of applicants applying to University and the universities in order to try to meet this increasing demand increase the number of admissions the irony here is they're increasing admissions but the admission rates are actually still declining because they can't increase the admissions at wait to meet the demand so not only are we getting this massive growth of students entering the universities it should actually theoretically be bigger than it is putting us at an even worse position and that leaves you with the question what can the university do to try to combat this increase in student population so as we saw one issue is if you want to try to keep the same class sizes you have to hire more faculty members and I'm sure if you talk to any faculty member they're going to tell you it's not a very easy process or you have to hire people you have to put some type of tenure track in plan and it's something that you just can't do rapidly enough to meet the the growing student population so the University resorts to these massive classes and I was discontent with how they executed this but it was actually interesting to me that they tried to use technology to help remedy this issue that they were having so in my exploration I wanted to see what other professions did we use technology to try to substitute and how can we try to learn from those successful things and apply them to education so throughout history we've had a good number of jobs that by now you don't really even go to a professional to use you have some app or some tool that can do it for you and you know taxi drives are still in business but I'm thinking in 2020 maybe not even those will be around anymore if Google can make self-driving cars and if you notice the one thing that all of these technologies have in common is that it takes this profession and tries to bring this directly into the hands of the consumer they try to cut out the middleman entirely so my thought was how can we somehow apply this to a classroom setting and hopefully replace this professor with some type of technology that brings education into the hands of the student themselves so first I'm going to talk about some critiques people had with online education and try to address those so one critique I've seen is that the traditional lectures provide the optimal form of learning and it turns out that this is actually online learning argument aside just a blatantly wrong statement in 1984 Benjamin bloom did an experiment where he tried to test different types of learning systems and saw that the traditional classroom setting was one of the least effective ways of disseminating knowledge to students and he said a quote that really resounded with me he said it makes no sense to expect all students to take the same amount of time to achieve the same goals objectives and this is intuitive just basically if we have a bunch of different students in the same classroom coming from different backgrounds how can you expect two different students to learn the same exact material at the same pace and what I've enjoyed about online education is that you can address this by breaking down the content into individual atomic units and then different students can take as long or as little time as they want on each unit another complaint that I've heard is that the classroom instruction indifference from MOOCs which are massive online open courses uniquely personalizes learning and when someone says a lecture personalized is learning all I can think of is are you serious there's absolutely nothing personalized about this at all so what I see education as is the stereotypical iceberg picture that all TED Talks have to have where we have the classroom lecture on the surface where this is basically where the students being exposed to a wide amount of information but not really diving deep into any of it and then the core pieces of their learning go on behind the scenes when the students actually trying to understand the material and as they're doing problem sets they'll encounter what we call learning breakdowns places where they just cannot get past some specific topic and it's through addressing these learning breakdowns that they're doing the bulk of their learning so ironically in the current system the professor is just involved at this stage which is really something that you can learn from the internet or from a YouTube video of some sort but the place where that the professor's expertise would be extra useful would be in addressing these learning breakdowns and we don't see any involvement from the professors in this stage so my proposition is trying to use online education to maybe flip this type of situation have the professors be involved with the learning breakdowns and not the general content so this triggered what I call the MOOC revolution where as I mentioned a MOOC is a massive open online course so the motivation of this movement was people saw it we have these massive class sizes and they're saying I personally wouldn't prefer this I don't know about you but this didn't work for me this is a little bit better this is a step in the right direction and want a few class would be pretty nice but ideally what I would want is some type of one-on-one interaction with the professor unfortunately as we mentioned this is unfeasible because we have such student bodies you can't have one professor to each student so these online courses are trying to use technology to hopefully simulate this one-on-one instruction by putting a professor trained device to help these students learn and this is where I come in so over the span of the past few years we've had this out surge of all these different platforms that people can create courses on and people have created courses that teach biology computer science math history pretty much any topic that you want to learn unfortunately even though these serve as a great reference they still in my opinion do not have any ability to substitute a real class and to kind of exemplify this I did an experiment of sample size one which was just me where I wanted to see Coursera this one of the main platforms of online learning had this new certificate program and I decided you know what this certificate seems meaningless to me I want to prove that it's actually a meaningless certificate so I enrolled in what's considered one of the best Stanford hosted online courses through Coursera which is the machine learning course and my goal was to finish this course to full completion without actually learning anything in the class and my solution to this was I just fast forwarded through all the lecture videos clicked massively on all the multiple-choice questions and then for the programming challenges I just used basic linear algebra knowledge to rearrange equations and eventually get the answer and I was able to successfully complete this 11 week course in five days getting 100% in the class without actually learning anything so I'm not saying that the course is bad if you want to learn machine learning you can learn it from this course but this certificate that you get when you complete it is clearly not a very meaningful thing and you can't have it replace the traditional diploma because a diploma needs to infer you know it has to imply some set of skills that you've mastered so in my opinion we're not at that stage yet that we can just replace online replace university classes with online learning and to kind of pull through with this discontent we have a famous mathematician he said if I had my wish I would wave a wand and make MOOCs disappear and I can see where his discontent comes because I have myself proved that not all MOOCs are created equal but I wanted to just draw a parallel with something that a famous scholar in 1492 said where he said the printed book made of paper is made of paper and like paper will quickly disappear so to scholars that had this negative opinion of some new way of disseminating knowledge clearly one of them was wrong because I still have a textbook for every one of my classes and I'm hoping that through refinement we can hopefully prove dr.

Vardi wrong as well and show that online classes can actually be successful and this is where I come in so my goals are to transform the impersonal MOOC experience into an emulated equivalent of a one-on-one professor student experience and I want to have students be able to address learning breakdowns even more efficiently than they normally be able to do even in a small classroom setting and this is where I transition from MOOCs to what I call ma I TS so m AIT is a massive adaptive interactive text so MOOCs in general are video based whereas I'm proposing an interactive text that you can use where so with traditional learning basically you have this linear sequence of information that you have to learn maybe the student understands the first step they keep understanding and eventually they hit what we call a learning breakdown and they just fundamentally do not understand this stage of the learning with traditional learning the student carries that misunderstanding all the way through and ruins their experience with anything that depends on this initial point so with our adaptive learning basically what we want to do is when the student encounters a learning breakdown we want to have some professor created detour that the student can take to address this learning breakdown specifically and then maybe while they're doing this detour they encounter another learning breakdown take another detour and address that and eventually hopefully all the dependencies will have been met so that they can return to the original content and this time understand it properly and then if they can do that then they'll be successful in the long run so we've implemented this type of a system in a number of online courses on the Left I show the bioinformatics algorithms series that I've been involved with it's been created by Pavel pepper and Phillip Campo and on the right I figured I would try to make my own equivalent of something like that so ITA the advanced data structures at class at UCSD and I didn't like the textbook and none of the students were reading it so I figured I would just make my own and over the span of a summer with the help of one of the undergraduate students we co-authored this interactive textbook that so far has stars on the host it's on has 2,500 students internationally and has roughly I think a thousand or so students within Eustace T that have used it so this stuff can actually work if you do it properly so how do you create a successful M AIT in my opinion you always have to start with some captivating attention grabber that's going to pull the student in something to just give a little bit of shock factor and make them initially interested then I think you have to motivate whatever topic you're teaching with some type of a real-world problem something that shows the student why exactly they're being forced to learn what they're learning then as you go through the content that you're teaching you should break down every single step of instruction into just a single message you shouldn't have more than one message or one takeaway per individual unit of topic you need to implement some form of active learning needed to have the student actually do problems while they're learning the material not after the matter because if they can do problems while they're learning the material they can instantly figure out where exactly they're learning breakdowns occurred and address them immediately instead of having all these fun down the line learning breakdowns that are a result of this initial one and then when you finish the topic you should always go back to that initial real-world problem that you started to motivate with and show that you've solved that successfully and hopefully leave the student with some lasting message that they can learn from whatever platform you hosted on it needs to have some mechanism of having students being able to discuss with one another and the kind of beauty of having these massive classes that you have students from all around the world with all different backgrounds able to help each other me I'm an American student that did a bio informatics undergrad and is doing a bioinformatics PhD I might not be able to think in the way that someone that did maybe a pure biological study would take to a given problem so maybe they're learning breakdowns that they encounter are just things that I can't even perceive so by having this massive class you better your chances of having other students have the exact same problems in helping each other out and of course you want to prevent people for me like me for me able to just kind of hack their way through your course so the most important thing as well is that you should make your course memorable and just in in general an enjoyable experience because students want to learn if you can make it seem as if it's fun to learn so why transition to M AIT specifically the first step it works right we did a survey on the Coursera massive courses where we saw that students significantly preferred the interactive text to the traditional lecture videos secondly it's a great time investment the investment pays off so I'm not going to lie it's a huge initial investment you have to increase your fixed cost substantially but then the variable costs associated with each subsequent iteration of the class is basically nothing you've already created all the material and he effectively just press play so Paul Lopez owner he's my PI he said I have not given a course lecture in over three years he created this initial content and he just basically has the students do the learning online at home and then during lecture time he spends the entire time addressing students learning breakdowns another important feature is that the open student accessibility in our courses we have students from all age groups various academic backgrounds from all over the world taking the class and most importantly to me you have student people that are full-time workers and parents with young children who are enabled now able to access this material so people who otherwise might have time commitments that make it difficult for them to pursue some form of higher education you've now opened the market of education to these individuals and it's a win win the student gets to pursue this education that they want and the university actually gets to open up the consumer base of their product their degrees and access parts of the market that they previously were not able to reach and on a kind of an extension of that it's financially sustainable so this is the John Hopkins University data science specialization and it initially had grant money to first be created but after just two months of execution they were self-sufficient entirely so it does pay off so just some concluding thoughts I have about this type of notion right now we're still at something like this most UCSD classes that I've been involved with are some massive lecture with one professor I'm hoping that with good practice we can eventually push ourselves to something like this where every single person all around the world of all different age groups can have their own many pawel to be able to learn from thank you

As found on Youtube