Right now, millions of people from all around the world use massive open online courses to get access to online education. The system's basic principles are very simple; as of now, MOOCs have already become one of the best formats for distance learning.
The OpenEdX platform has many such courses. It already contains thousands of educational programs developed by different educational establishments. Previously there was no interconnection between them - no direct links from one course to another that would have helped the user get comprehensive knowledge in a particular field.
Such an approach is quite effective, as the users don't have to search for anything by themselves, yet it doesn't really work well when one looks for an answer to a particular question. This was the problem ITMO's scientists worked on.
To solve it, the PhD students created an aggregation system based on the ECOLE semantic technology. The system already offered many interesting solutions, yet it did not factor in the latest trends in the field of massive open online courses.
"The main problem was that initially, the system worked by itself, and was not bound to any platform. It is now extremely widespread, many Internet users work with it, so we've decided to develop it further," explains Dmitriy Volchek, PhD student at ITMO's Department of Informatics and Applied Mathematics.
The researchers focused on semantic representation of MOOC content. As a result, ISST laboratory's staff and specialists from Yandex have created an ontological model that describes all of a MOOC's aspects, and developed it in a form of a software plug-in.
"We are talking about a simple installation mechanism for those who use the OpenEdX system; it can be installed as an add-on using the Xblock SDK. One can use it just like any other program we install on our computers. All you have to do is write several entries on the server," shares Dmitriy Volchek.
Thanks to this solution, semantic technologies have been introduced into the educational process. The primary goal of the project was to reveal the interconnections of the platform's courses and thus create interdisciplinary links between them.
Another advantage of semantic technologies is that they allow for representation of data in a way that is comprehensible to both humans and machines. The developers plan to structure all the MOOC content according to their new ontology; in layman's terms, that would be explaining what every part of each course is about and what its components are.
"We use the following NLP algorithms: take the content represented as text (subtitles in case of video lectures) and derive the key words - we call them "target field concepts", which we tag for ontology. Thus, we get units of meaning for each course, and can use them to interconnect different courses and track the use of concepts in different fields of knowledge. We also monitor the quality of the content: for instance, we analyze whether the theoretical materials provide enough knowledge for the users to complete tests – in other words, whether the concepts in the texts correspond to the tests," explains Dmitry Volchek.
Let's say that you're doing a course in mathematics and learn a new term –“convergence”. The new system allows you to learn where else this term is used, in other courses or fields of knowledge - physics, for example. The system works differently for those who use courses and those who create them; the latter gain access to links where they can see where else a particular concept is used so as to get access to courses from similar fields of knowledge.
"First of all, we solved the problem of adapting the system for the current OpenEdX platform, which makes use of a different server, technical and program solutions. The second task we worked on was the introduction of semantic technologies to the field of education. Many research either education or semantics, but we've completed a project that unites these two fields," concludes Dmitriy Volchek.
As of now, the system is in its development stage; the researchers are currently testing it. The first results were presented at the OpenEdX conference and got positive feedback.
In future the developers plan to track the users' individual educational paths so as to better adapt their system for each particular user's needs.
