How semantic technologies improve online education: a project by ITMO University scientists
You can study with ITMO University faculty without admission - just sign up for a massive online course. The University produces such courses not only on domestic platforms (such as Open Education), but also on the edX platform, which was created by MIT and Harvard in 2012.
For example, last year, the edX platform launched the English-language course “How to Win Coding Competitions: Secrets of Champions” (we already talked about it earlier: 1 , 2) However, ITMO University not only creates interesting courses, but also seeks to improve the work of the educational platforms themselves. Scientists from ITMO University together with a colleague from Yandex LLC have developed a solution that allows the introduction of semantic technologies in the educational process on edX.
/ Photo by Craig Sunter / CC The
ability to analyze data from thousands of students is one of the benefits of massive online courses. However, a number of difficulties are also associated with mass character - platforms need to take into account more and more requirements from their diverse audience. Some of the issues that arise when creating courses and working with students are solved by the community of developers, teachers and scientists designing and conducting courses - for this, edX has an open source platform Open edX. It was precisely its capabilities that our researchers used in their work.
Work on the project was carried out by scientists from the international laboratory Information Technology and Semantic Technologies (ISST) of ITMO University and their colleague from Yandex. In Russia, ISST is one of the leading teams in the field of semantic technologies and ontological engineering.
Until recently, courses on educational platforms were perceived as an “indivisible complex of knowledge” - in order to understand some narrow issue that is addressed in the course, you must complete the entire course (or at least get to the necessary lecture). In this regard, “interdisciplinary listeners” had a hard time: in this format, it is difficult to look for answers to specific questions that are covered only in a small part of a given course. And all the more difficult is it to study the issue that is being considered in several related fields (and, therefore, in several unrelated courses).
/ The structure of the online course. Material from the presentation“Metadata Extraction from Open edX Online Courses Using Dynamic Mapping of NoSQL Queries” (Dmitry Muromtsev, Alexey Romanov, Dmitry Volchek, Fedor Kozlov)
The task of scientists from ITMO University was to help students and create a solution that not only allows you to create a “custom set of courses” in accordance with the question of interest, but also to do so taking into account all the technological features of the modern educational platform.
Scientists solved the problem of the semantic presentation of the educational content of the online course. As a result, they managed to create a model that describes all the characteristics of the course: its content, use cases, participants, etc. This approach allows you to “parse” the course into components, describe and structure its content in accordance with the developed ontology.
According to Dmitry Volchek, graduate student of the departmentIT and ITMO University informatics and applied mathematics, the algorithm analyzes the course content (text materials, video lecture subtitles), selects keywords (“domain concepts”), after which the concepts are marked on the prepared ontology. As a result, it becomes possible to show how courses from different disciplines are related, find common topics in them, and check whether the selected content is reused in other lectures and materials.
This approach allows the creators of educational materials to evaluate the quality of the course: a comparison of concepts in tests and lecture materials gives an idea of whether the level of control problems corresponds to the volume of theory that was issued at the lectures. And students have the opportunity to find courses from related fields or take a fresh look at the question under study: for example, to understand how a particular concept is considered within the framework of various disciplines.
/ Ontological model (course and users). Material from the presentation "Metadata Extraction from Open edX Online Courses Using Dynamic Mapping of NoSQL Queries" (Dmitry Muromtsev, Alexey Romanov, Dmitry Volchek, Fedor Kozlov)
Now the project is at the testing stage, however, scientists have already presented the first results at the Open EdX Conference 2017 profile conference (the conference was held in Madrid this May and brought together developers, analysts and educational specialists working with the edX platform). In addition, according to the results of research, scientific papers have already been published ( 1 , 2 ).
For example, last year, the edX platform launched the English-language course “How to Win Coding Competitions: Secrets of Champions” (we already talked about it earlier: 1 , 2) However, ITMO University not only creates interesting courses, but also seeks to improve the work of the educational platforms themselves. Scientists from ITMO University together with a colleague from Yandex LLC have developed a solution that allows the introduction of semantic technologies in the educational process on edX.
/ Photo by Craig Sunter / CC Theability to analyze data from thousands of students is one of the benefits of massive online courses. However, a number of difficulties are also associated with mass character - platforms need to take into account more and more requirements from their diverse audience. Some of the issues that arise when creating courses and working with students are solved by the community of developers, teachers and scientists designing and conducting courses - for this, edX has an open source platform Open edX. It was precisely its capabilities that our researchers used in their work.
The project team
Work on the project was carried out by scientists from the international laboratory Information Technology and Semantic Technologies (ISST) of ITMO University and their colleague from Yandex. In Russia, ISST is one of the leading teams in the field of semantic technologies and ontological engineering.
Task
Until recently, courses on educational platforms were perceived as an “indivisible complex of knowledge” - in order to understand some narrow issue that is addressed in the course, you must complete the entire course (or at least get to the necessary lecture). In this regard, “interdisciplinary listeners” had a hard time: in this format, it is difficult to look for answers to specific questions that are covered only in a small part of a given course. And all the more difficult is it to study the issue that is being considered in several related fields (and, therefore, in several unrelated courses).
/ The structure of the online course. Material from the presentation“Metadata Extraction from Open edX Online Courses Using Dynamic Mapping of NoSQL Queries” (Dmitry Muromtsev, Alexey Romanov, Dmitry Volchek, Fedor Kozlov)
The task of scientists from ITMO University was to help students and create a solution that not only allows you to create a “custom set of courses” in accordance with the question of interest, but also to do so taking into account all the technological features of the modern educational platform.
Decision
Scientists solved the problem of the semantic presentation of the educational content of the online course. As a result, they managed to create a model that describes all the characteristics of the course: its content, use cases, participants, etc. This approach allows you to “parse” the course into components, describe and structure its content in accordance with the developed ontology.
According to Dmitry Volchek, graduate student of the departmentIT and ITMO University informatics and applied mathematics, the algorithm analyzes the course content (text materials, video lecture subtitles), selects keywords (“domain concepts”), after which the concepts are marked on the prepared ontology. As a result, it becomes possible to show how courses from different disciplines are related, find common topics in them, and check whether the selected content is reused in other lectures and materials.
This approach allows the creators of educational materials to evaluate the quality of the course: a comparison of concepts in tests and lecture materials gives an idea of whether the level of control problems corresponds to the volume of theory that was issued at the lectures. And students have the opportunity to find courses from related fields or take a fresh look at the question under study: for example, to understand how a particular concept is considered within the framework of various disciplines.
/ Ontological model (course and users). Material from the presentation "Metadata Extraction from Open edX Online Courses Using Dynamic Mapping of NoSQL Queries" (Dmitry Muromtsev, Alexey Romanov, Dmitry Volchek, Fedor Kozlov)
Now the project is at the testing stage, however, scientists have already presented the first results at the Open EdX Conference 2017 profile conference (the conference was held in Madrid this May and brought together developers, analysts and educational specialists working with the edX platform). In addition, according to the results of research, scientific papers have already been published ( 1 , 2 ).