You are here :
Traces d'apprentissage: big data and artificial intelligence to help learners succeed
With Kosmos next, its research and innovation program, Kosmos has chosen to surround itself with a team of experts to reflect on various subjects. The aim is to improve Student success and build the school of tomorrow, using new technologies and the immense range of possibilities they open up.
Learning Traces: Big Data for Academic Success
Today, it's possible to generate a wealth of data known as learning traces. These Traces are derived from the use of digital tools or Online services, and make it possible to gather information on the achievements and learning behaviors of Students. an important issue for the future of education.
Learning Analytics or Learning Analytics is "the discipline devoted to the measurement, collection, analysis and reporting based on data from learners in a learning context with the aim of understanding and optimizing learning and context".
The field of Learning Analytics derives from Data Analytics as applied to the education sector, and includes the concepts of Data Mining and Machine Learning.
Digital environments such as ENT or EMS (Education Management Software) used in Schools are ideally suited to the collection of learning traces, due to the sheer volume of data collected. EMS are capitalizable platforms with over 50 million visits per month.
Kosmos on the road to Adaptive Learning thanks to big data
Through its partnership with Loria (Lorraine research laboratory in computer science and its applications) and a joint project to collect and analyze learning Traces from SkolengoEMS (Education Management Software), Kosmos is taking part in discussions on these innovative subjects.Data from Skolengo's Exerciser and Teaching Services will be stored in Learning Record Stores (LRS), currently based on the xAPI standard (reference standard for processing learning traces). In this context, Kosmos guarantees data collection and use strictly compliant with the RGPD.
The next step will be to work with our experts and partners to determine how best to use this information. The difficulty of analyzing this data is akin to the issues surrounding artificial intelligence in a Big Data context: from a substantial volume of seemingly uncorrelated data, how can we extract relevant information?
The project, entitled METAL, brings together players from the education sector, schools, universities, research laboratories and Local authorities, and proposes to. "design, develop and Grade a set of individualized monitoring tools for students or teachers (learning analytics) and innovative technologies for personalized language learning".
Although Adaptive Learning is still in its infancy, the subject of this study suggests great potential for imagining individual and 100% adapted Teaching paths.
Based on the same principles as recommender technologies, Adaptive Learning could enable students to be offered personalized learning paths, thanks to the analysis of data already collected on previous Pieces of work. This approach borrows many features from similar content suggestion tools. Just as Google, Amazon and YouTube harvest data to identify their users' preferences, Adaptive Learning could enable data to be stored in order to propose content tailored to each individual to improve their performance and thus take into account the uniqueness of each learner.
In the future, these practices could be automated to easily generate individualized learning paths for students. But it will also provide invaluable support for teachers and administrators in identifying profiles requiring special attention and assistance, so as to provide the best possible support for all students in their teaching careers.