Introduction to EdTech

The education industry is a huge, multi-billion dollar industry. And as an industry concerned primarily with imparting information, knowledge and skills, it’s ripe for disruption. After all, the Internet democratises information by making it available to everyone with a connection fast enough to access it. And typically, industries with large inefficiencies that can be rectified by faster and better communication of information, such as the music and video industry (Spotify and Netflix are the kings now, not physical CDs, DVDs and Blurays) tend to lend themselves well to technology-based disruption.

Thus we find ourselves watching a thousand and one startups in the burgeoning field of EdTech (Education Technology), trying their luck at cracking the code to technology-enhanced education.

This article concerns itself with explaining the various approaches to EdTech out there, and what they are good and bad at doing. Some of them seek to replace education altogether, and some aim to complement education in schools. At the end, I end off with some observations on what areas of education aren’t, at present, covered by technological solutions.

Massive Open Online Courses (MOOCs) AKA Lecture Videos

Probably the easiest idea to come up with and implement, this one, at the core, entails getting a lecturer to take a video of himself/herself and upload it for all to be instructed. What began as YouTube videos made by hobbyists has gone legit, withstartups like Coursera, EdX, Udemy, Udacity courting universities as great as Harvard, MIT, Stanford, Yale, Cornell, Duke, UMich, Northwestern, etc. to get their professors to create lecture videos as part of a series that aims to develop a certain skill. Some may offer automated marking of work too, allowing you to gain feedback. For example, you could submit your code for an assignment, where it will be tested for accuracy and speed. Sometimes, there are startups such as ExecOnline that deal directly with educating professionals; these make deals with organizations to educate their staff, which is appealing to our corporate overlords because the courses can be taken online, reducing downtime of the staff.

Interestingly, the skills focused on tend to be programming-related. On Coursera, there are tons of lectures organized into courses, with some courses even grouped into specialisations (basically a group of related courses that build up an overall ability). A lot of them are on coding languages, Data Science, Algorithms (in Computer Science), Machine Learning, etc. Granted, software programming is the “in” thing to learn now because it’s profitable, but I feel that there’s too much focus on this, and very little on other subjects that are also useful like Economics. And besides, education is not just about learning useful skills, but about enriching the soul. I’d love to delve deeper into Philosophy, but I highly doubt these EdTech startups are going to care about such subjects for a long time. Lastly, perhaps the reason software features so heavily in these courses is because the startup founders are usually coders, which means that they’d likely prioritize their own field and hence seek to have more software-related courses. You can’t really expect most Engineers to give a crap about Literature, can you?

Overall, I think these are a good first step, but they lack in some areas other than variety of subjects, which can be solved by getting more professors to create the required courses. One area that lecture videos cannot get right is interaction between student and teacher, which is necessary for good feedback to be gotten. You usually can discuss with other learners, but then again, you’re discussing with other learners. Another is that lecture videos reduce education to the acquisition of skills. This will never ever replace higher education, which is usually as much about finding yourself and going beyond your boundaries as it is about learning skills that will feed, clothe and shelter you in the years to come.

Still, it does really well in allowing you to learn at your own pace. You can sometimes speed up lecture videos, or go to the next video if you have the time and energy to. This means that that month-long course can be completed in under a week if you dedicate yourself to it. And if you’re not as fast a learner, you can slow the video down and ensure you understand everything before moving on. A lecture in real life would leave you lost if you can’t keep up.

It’s also much cheaper than paying a professor to teach you in real life, because these videos would generate passive income for them, meaning that once the videos exist, the professor or teacher earns from their one-time effort. When the returns can be made across a long period of time and thus a large number of students, the cost per student is lowered dramatically. In fact, all Coursera courses are free to be audited, except for getting assignments marked or gaining a certificate certifying you have completed the course. It costs only about a hundred or so per course though, which is still quite cheap considering you’re getting professors at the best universities like Duke to teach you.

Education Platforms/Portals/Course Management Platforms

These have actually been around for a while, and work directly with schools. Examples include Blackboard, and Moodles.

Basically these are centralized platforms or social networks (somewhat!) that allow for content to be uploaded, such as lecture slides, homework, etc. Students can also hand in work here. Teachers can al
so post updates for all students to read. They do well in terms of ensuring that all students have access to material. It’s much better than emailing everyone a file and then struggling to keep everyone up to date on the latest version of the file. Students can also ask teachers questions that everyone else can see. I’m sure the primary use of this would be for homework extensions, so everyone can get proof that the teacher has extended the deadline. (On such nights, I’m also quite sure that little work will be done after the teacher has extended the deadline, making one wonder why students can’t get their shit together.)

I’m not sure if there’s much space for growth in this arena, other than getting teachers to actually use them more frequently for interacting with the class. These websites can be useful, but they certainly don’t do very much other than keeping everyone on the same page regarding information and documents.

Plagiarism Detectors

Nightmare of lazy students everywhere, plagiarism detectors like TurnItIn have been around for a while. They run automatically on each piece of work you submit through an education portal (assuming the education portal integrates it into their site) or sometimes, teachers directly run your work through the checker itself. What they do is simple: they compare your work to a database of work that’s been uploaded, and check for similarities. How they do it, I honestly have no clue. It’s proprietary stuff, and it works just as much by being successful in checking for plagiarism, as by making students scared enough to submit wholly original pieces of work.

Still, I doubt that it would consider these two sentences to be similar to each other:

  1. The quick brown fox jumps over the lazy dog sleeping on the floor.
  2. The fast, chocolate-coloured fox leaps across the sluggish canine dozing on the ground.

We know that they give essentially similar information, but a software that compares similar words would only find some similarities such as the word “fox”, and “the”. Everything else is different. If the software is able to compare sentence structure, it would see more similarities, but then would it be fooled by this?

The sluggish canine was dozing on the ground when the fast, chocolate-coloured fox leaped across it.

In isolation, these changes may not seem like much, but what if every sentence in some essay is tweaked in such a way? Overall, the content would be the same. One could even preserve the entirety of the line of logic. Yet the software may not catch it, unless it’s able to understand meaning. Language is more than syntax, it’s semantics too.

I believe these softwares can become much better at “perceiving” meaning as we learn to apply Machine Learning/Neural Nets/some as yet unheard of Artificial Intelligence method to them.

Automated Grading

In some way, this is similar to plagiarism detectors. Either you submit a piece of work online and the software automatically grades it, or your teacher directly runs your work through the software which grades it. And they also have the same pitfall of not understanding meaning, as plagiarism checkers.

We already use automated grading called Optical Mark Recognition (OMR) for multiple choice questions with Optical Answer Sheets, otherwise known as OAS. These are great because they save time on grading, which allows teachers to have more time to focus on one-to-one instruction, or creating notes or doing something where their skills can be put to use. Also, automated grading allows for teachers to automatically collate data on what sorts of questions students aren’t able to do, which allows them to focus on weak areas that classes generally struggle in. All in all, this allows for an education with more teacher-student time, and more personalized learning.

But they work only for quantitative subjects like Math and Science, and even then only partially. We can’t get software to grade just how well a student explains Newtonian Mechanics or the Theory of Relativity. Perhaps he or she got the right answer but wrong working? Or the right working but accidentally made a careless mistake?

And for now, there’s no way we can judge the quality of prose written by a student. There are some that are attempting to create Algorithmic Essay-Grading, but results so far aren’t great. I’m a firm believer that writing is a science, not an art; in fact, I think that all art is a science that we have found so complex that we can only go by intuition and some principles we have observed. That’s precisely the problem though. How can we lay down the rules that underlie the infinite dimensions of the written word? We simply cannot. This takes the problem faced by plagiarism detectors one step further. Even if our software can identify meaning, could it judge the usage of metaphors and other literary devices, or ascertain both the originality and quality of a simile? It’s so complex that we ultimately just go by “feel” or gut instinct, but clearly there are reasons why one piece of writing works and another doesn’t. Neural Nets may solve this problem, but until then we’re going to have to wait a long time.

Still, this technology has a lot of potential to allow for one-to-one instruction, which is usually a good thing. If we can even get it to work.

Gamified Learning

Basically, learning through games. You’ve seen some of these in primary school, I think, but they’re really trashy and lousy games. These use the interactivity of software and/or hardware to make learning more engaging and fun, which do more than make students spend more time on the material. They increase retention rate, inspire and motivate them to learn even more and become the next generation of scientists, etc.

Nowadays, there is software like Duolingo which makes learning another language fun because it condenses learning into bite-sized topics and allows you to play them again to get a higher strength bar or a higher score. Still, on the topic of Duolingo, it doesn’t address a lot of the problems of learning a language, but I shan’t go into that now as this article is already ridiculously long, and it’s hardly my focus here.

Sometimes, learning is made competitive, so players try to outdo one another. I feel it works best if you learn within a community of friends, so that you’re actually incentivized to compete. Otherwise, I don’t really care that KrazyL34rner has 6 billion times my score.

There’s tremendous potential here when it comes to self-learning and for getting adolescents to learn. Self-learning because it’s not always easy to motivate yourself when there’s always so many other things to do, and adolescent learning because they are likely to be less able to pay attention and being engaged through a game would motivate them more.

I believe that with the rise of Virtual Reality (think: Oculus Rift, Google Cardboard), Augmented Reality (think: Microsoft Hololens) and Mixed Reality (think: Magic Leap, which isn’t even available to the public yet) systems, learning can be made much more engaging and interactive. For now, it’s gonna be really expensive too though, but I’m hoping that it’s going to be the future of education, and go beyond merely teaching theory, to teach practical skills too. Imagine having a piano that your AR/MR headset overlays with information like the key, which may be helpful for beginners. That’s up to the teachers to decide, but there’s a lot of potential here.

Memorisation Tools

These aim to incorporate memorisation methods into the learning process. Some methods are Spaced Repetition (learning over a period of time with increasing intervals between revisions) instead of Massed Repetition (cramming Japanese during an immersion trip to Japan, or cramming the night before, which drastically improves skill which is also lost very quickly); or mnemonic memorisation techniques like the Mnemonic Major System, the Method of Loci, or visualization techniques. There are other general practices that learning scientists will know about, and these combined can make learning a lot more effective. The problem is implementing them, which is where software comes into play.

There are a few examples. Anki SRS (Spaced Repetition Software) uses a virtual flashcard system to quiz you on things you are trying to learn. It’s very irritating to use because the app isn’t designed nicely and it’s really difficult to set it up. Memrise does the same but crowdsources their flashcards so it’s much easier to learn from it. Vocabulary.com has a database of words with questions and allows you to learn from them, which is the approach I prefer because you are ensured that what you learn is legitimate and standardised. Of course, they have to hire a team of professionals to create quizzes, but overall it’s a better learning experience than, say, Memrise, which has a lot more breadth but just isn’t as good in terms of quality.

There’s a lot of potential in getting schools to utilize Spaced Repetition Software, I feel, and getting students to use it as part of their daily homework. It can make learning more fun, and certainly makes it more effective.

One problem that needs to be solved is that beyond quizzing students on Multiple Choice Questions, they need to be able to get students to give answers entirely on their own, which is tougher than being prompted. For example, Vocabulary.com has improved my reading comprehension, but it does not help me to learn to use highfalutin or sesquipedalian words in everyday conversation. To summon these words of your own accord takes more practice and a different type of learning than being able to understand what others write.

Unique Approaches I Cannot Categorise

One of these is FluentU, which is very innovative.

What it does is get teams of translators to transcribe and translate videos so that as you watch videos, you can also learn another language. This is great, because it exposes you to so much more than say Duolingo does. Duolingo is largely text-based and mostly neglects understanding accents and comprehending speech longer than mere sentences. Videos are a truer representation of real-life speech, and thus one can learn a lot more by trying to understand what a foreign speaker is saying. Yet it’s difficult to even hear what they’re saying, so the transcriptions help with the initial learning stage of figuring out what they’re saying. One can also watch a video in their home language while seeing a translation in the subtitles. Lastly, words that one doesn’t understand can be added to a word bank, to learn later. It’s highly innovative and could really help someone learn another language in a fun and useful way.

Last but not least (in fact one of the most interesting of them all, in my opinion), Minerva KGI, which is in a class of its own for its approach in EdTech.

Minerva KGI is actually a for-profit university (don’t get turned off yet) that aims to use technology to bolster learning. It does away with a traditional campus, using the city as a model for the campus, and eschews typical lectures for small online video chat-based classes with a professor and about 20 other students. These classes facilitate student-teacher interaction and full participation in class as students have to give their responses to questions and engage with the material in real time. It costs a lot less than typical American universities because it doesn’t have much in the way of facilities. And you get to travel around the world to 7 different locations, as you take 2 semesters in San Francisco, and a semester each in Berlin, Buenos Aires, Hyderabad, London, Seoul and Taipei. It’s supposedly extremely selective, and does away with SATs and other standardised testing in favour of its own battery of tests to analyse critical thinking ability and aptitude. The downside is that it’s so new, and unproven; the first batch hasn’t even graduated yet. Considering that this requires a huge time investment, and you will be doing your bachelor’s here, it’s a huge risk. That said, it is a very interesting start-up that I’m keeping my eye on.

Conclusion

There we have it, an overview of most of the different areas in EdTech today, and a peek at the future to come. I’m sure there will be many more brilliant ideas and I’m excited to see learning transformed by technology’s disruptive power. If there are any other categories of EdTech approaches you are aware of, do let me know and I’ll update the article.

Leave a Reply