Scroll down

Smart Tutoring: the smart tutor which helps students to progress individually according to their learning needs

Benoit Praly |

The health crisis and technological advances transform teachers’ educational practices. Digital uses in class and at distance tend to accelerate, leading the way to new perspectives: some with positive results, as time-consuming tasks automation and data collection, and others a bit less like digital excess and tools abundance. This particular time is yet still likely to bring hope and opportunities if we succeed in creating digital tools that exploit learning data, making work easier for teachers as those tools would be integrated into their pedagogical practices.

Being in charge of 30 students while having to follow-up individually each of them seems hard. Having a good knowledge about their learning needs, in addition to in-class teaching and others time-consuming tasks outside the class seems almost impossible to achieve. Domoscio offers to teachers to be assisted by a “smart tutor”, whose purpose is to understand, thanks to learning data, the learning needs and cognitive profiles of each student to offer them automatically a differentiated and personalized learning journey.

smart tutor

1 – Assisting teacher to enhance his students’ comprehension

Learning paths at school are most of the time based on a global approach. This means that learning journeys are identical for all, regardless of pedagogical objectives, level of mastery, and each student’s ability to progress. The textbook is a good illustration of this approach: a learning material that can be opened from page one to the last but have no adaptivity at all. When using it, a student will learn from pedagogical contents that are identical for all, and none of them can be assured to be prevented from comprehension trouble. On the opposite, the smart tutor will learn from each student learning needs to suggest him adapted learning paths and help him overcome potential comprehension trouble in order to better progress.

This individualization can be done “manually” by the teacher himself but requires a lot of his time. The learning path’s individualization approach would then benefit greatly from taking advantage of digital tools.

The smart tutor purpose is to understand the students’ needs and learning profiles to individually and automatically provide them with adapted and differentiated paths. The use of an adaptive learning technology comes up with learning activities’ suggestions (practical exercises, lessons, videos, methods, etc.) that will allow a progressive and individualized upskilling. This will encourage students’ autonomy in their progression as they are guided by a tutor that suggests them learning activities, always adapted to their current level of mastery and expected one. This smart assistant is by the way a component of Domoscio Spark: a technological, innovative and modular product.

There are many application cases and uses associated to Smart Tutoring. It is possible to set up a connection between the recommendation engine and the student, who then finds himself independant and assisted in his progression. Other scenarios do exist though. For instance, when a specific context allows it, it is possible that learning activities suggestions are pushed to the teacher, who then turns out to be the mediator and choose which learning activities to assign to the student. In this case, the teacher remains the only master of his pedagogical choices in addition to benefit from indicators that allow him to follow his students’ progression.

2 – How to put into practice a learning individualization approach?

Adaptive learning is, -as science describes it- a recommendation system. Domoscio’s algorithms take advantage of a “collaborative filtering” model to produce the necessary calculations to make sure recommendation production works. This approach main benefit is to offer student learning activities and then be able to measure their impact on his progression.
The system updates itself and will readjust the recommendations level of difficulty for a given student, but also for the next ones whose profile and path are similar compared to him. Collected and exploited data by the model comes from:

  • Learners’ interactions with the suggested content (viewing, time spent, results, etc.)
  • The level of mastery/target skill’s validation phase
  • The mistakes made by students when they do their exercises
  • The students or teacher qualitative opinion on the interest and practicality on resources
  • Choices made by teachers or students when they are facing several learning activities recommendations or when they choose a specific activity on their own

Data collection benefits from a model enrichment that makes personalized learning activities recommendations. The same model can also be used on each learning activity and all students’ interactions. This tool can then be used for continuous improvement and curation of pedagogical activities hosted by e-learning platforms. In order to guarantee model’s transparency and explicability, functionalities integrating AI will provide a confidence rate to inform the user about the reliability of the recommendations.

3 – Already a success story

Domoscio, with the help of education players as Hachette Education and Beneylu is taking part in French Public Innovation partnerships on artificial intelligence and had the opportunity to lead an experiment and offer teachers its smart tutor called “Navi” to help them teaching reading and writing skills for Kindergarten and Elementary school (until 3rd Grade).

Navi is a teacher’s assistant. Thanks to AI and Domoscio’s algorithms, Navi suggests remediation material to teachers adapted to students’ abilities. Artificial intelligence individualizes the students’ learning experience through differentiated learning paths, avoiding then information overload for the student. Remediation in Navi allows to guide the teacher in order to identify resources (learning activities, practical exercises or lessons) adapted to each student to help them better achieve their skills development objectives. Nowadays, remediation rises enthusiastic reactions from teachers using Navi: “I am having difficulties implementing remediation progressively for each student. I couldn’t do it with them individually before using Navi” said a teacher taking part of the experiment. Currently at the end of its experimental phase, Navi project will be deployed on a large scale at the beginning of September 2021 in order to bring remediation’s benefits to the greatest number of teachers and students in France.

Benefits from Smart Tutoring have been proven. These benefits are even greater when associated to relevant data visualization for teachers and Schools: a “Learning Analytics” functionality that will be discussed on a coming article.

1 Pedagogical remediation is a means of addressing learning gaps. It is a way to remedy a learning situation where the student has not assimilated the content.