AI has long to be thought of as a technology that could automate predictable tasks that could be use in a commercialized setting. Well, those days are gone and we are at a time now where AI has become broader than that, specializing in many different industries and helping revolutionize the way those industries operate.
Today we will be exploring the current adoption of AI in education and its anticipated future.
In a Virtual and Physical Learning Environment
During the COVID-19 Fiesta. We saw majority of schools globally, utilizing virtual learning environments. Be it using Zoom or Teams to host lessons and a Learning Management System(LMS) to distribute assignments and initiate an asynchronous learning experience.
Uses of a Learning Management System(LMS)
With the rise of use in a LMS, lots of schools has started to realize the potential of online education.
Assigning assignments and tracking student's activity in a LMS. A teacher can utilize those logged data to make assumptions on each of their students understanding, and using those assumptions to reinforce engagement and support their student's developmental journey towards their own education.
Where the AI comes into place
based on the aforementioned data, there exist a lot of models that can interpret those data, and make useful classification, clustering, prediction and ETC. Tons of published papers had made initiatives to create systems designed to improve or revolutionize our approach to the traditional educational system.
Introducing Educational Data Mining(EDM) and intervention strategies
Before i go on a tangent on this topic, please keep in mind that the terms I will be throwing around are under the EDM field, and I wouldn't expect it to be grasped just by reading this published paper.
A lot of papers that are published in this field, deals a lot with metacognition and different subsets of it when it comes to education.
A common misconception of this field is that, its a statistical field. However there are systems being build that incorporates ML models and integrates this field and the technology that came out of it into learning environments. Blurring the lines between research and practice. The results of such application are still up to debate, and while many papers report of results and findings that are exceeding expectations, there are yet to be widespread.
Getting into the nitty and gritty of EDM
Referring back to the venn diagram above, each of the subsets represent the challenges and the features of a system that is needed to build such an architecture for a new educational system that utilizes technology and AI to its fullest. Of course for any other AI application in other sectors they require data to even make such a model.
Item Response Theory
Another AI model that has already been developed and integrated into some learning environment, are ATI's item response theory for student assessment on their growth and ability to certain subjects, which can be use to correctly and efficiently intervene individual students. Gone are the days of only plotting a bell curve distribution to assess a cohort.
Deep Learning models in Deep Knowledge Tracing(DKT)
A team at Stanford and in collaboration with Khan Academy and Google, has develop a Recurrent Neural Network Model(RNN) and a LSTM model to predict student responses to exercises based upon their past activity. Compared to a Bayesian model, theirs did significantly better. However it does come with a few challenges and some key factors missing(algorithmic bias), but shows that AI can be used in a learning environment alongside many other models for its different purposes.
What would this look like for students?
If you're a student reading this, expect more support coming from your teachers as they're able to understand you better. Unique assignments that is curated based on your engagement, instead of one learning path for all. Interactive educational contents based on individuals; Just like the online social media platforms you use that are designed to keep you engaged, by learning about your preferences, can also be integrated with an LMS that would recommend contents that suits your learning pace. Lastly, a totally cool learning environment, that you can totally rock in, and not worry much!
TL:DR; a better personalised environment(physical and virtual) will be designed to make education more intuitive and engaging.
What would this look like for teachers?
okay, I'm not a teacher. But I can also empathize with them, as they're also adopting the new learning models alongside their students
New Hybrid learning environment
High and tertiary education institutions in Singapore, has already adopted a hybrid learning environment way before the transition to a fully virtual environment due to covid-19. Setting a (E-Learning) day for students to be familiar with LMS i.e BlackBoard or Canvas, and to also utilize them during physical lessons. What this means, is that students that has already been familiar with the LMS environment are able to seamlessly transition into a fully virtual environment. However, as the covid-19 situation will inevitably be mitigated, we shall see new learning environments that utilizes both the physical and virtual environment pros. Alternating Blended learning model, is probably the correct summarization of what's being planned for the future of Singapore's educational system.
Probably another mandatory course about student data analytics
"Hais, another course?", is probably what you would say to yourself. Jokes aside, the intrinsic value of such course, is fundamental to the development of students. Leveraging the knowledge gain in this course can be the difference between a students understanding of a subject and none at all.
