— — Product – bettermarks EN

ADAPTIVE LEARNING SYSTEM FOR MATHEMATICS

LEARNING FROM MISTAKES, AHA EFFECT GUARANTEED

Titanom Technologies GmbH is proud of its long-standing and successful cooperation with bettermarks GmbH. In the course of this partnership, numerous features for bettermarks’ adaptive learning system have been jointly developed and successfully rolled out to hundreds of thousands of users. In addition, we researched various innovative use cases that will help to further improve the digital education of tomorrow. This cooperation has further consolidated bettermarks’ position as a leading learning system in the field of mathematics.

bettermarks GmbH

Bettermarks is the leading ITS (Intelligent Tutorial System) for mathematics in Germany and offers two applications to provide teachers with targeted support. “Practice & Test” includes adaptive tasks in which students receive comprehension-oriented feedback, while teachers receive clear evaluations. “Teach” offers interactive blackboard images for starting lessons, examples and mnemonics to reinforce what has been learned, adaptive tasks to consolidate practice, open tasks for higher levels of difficulty and tests to check performance. Driven by the vision of making learning easier, bettermarks decided in 2021 to work in partnership with Titanom to develop AI-supported, adaptive educational media.

Adaptive educational media for school lessons

Pupils are often unable to learn school mathematics independently – this is the subject with the greatest need for support. To enable them to recognize and correct their own mistakes, bettermarks has developed an intelligent tutorial system. This system is able to identify over 2000 error patterns and provide appropriate feedback. It recognizes gaps in knowledge and suggests fully automated series of exercises – from a pool of almost 200,000 different tasks that specifically train the learner’s weak points. In this way, pupils can learn independently from their mistakes and don’t get stuck in one place.

THIS IS THE PRODUCT

In the course of researching effective learning application scenarios for artificial intelligence, bettermarks GmbH and Titanom Technologies GmbH have initiated various research projects over the years. These included the analysis of feedback and analysis methods for free text input, AI-based task generation to speed up editorial processes, competence-based learning level measurements, the enrichment of metadata for individual tasks based on end user usage data and much more.

The results of this research work were integrated into the platform and rolled out to numerous users in collaboration with the teams at Titanom Technologies GmbH and the existing teams at bettermarks GmbH. To this end, prototypes were quickly developed, tested with user groups, adapted and then transferred to the existing infrastructure.

Analysis of free text entries: In addition to the rule-based evaluation of student input in the context of numbers and geometric shapes, new possibilities for task types and subjects arise as soon as free text input can be validated. The technologies and didactic principles required for this were designed and implemented in collaborative work.

AI-based task generation: Manual programming effort is required to create and parameterize adaptive exercises in order to formally define the logic within the exercises. This effort represents a significant part of the editing process and can be made more efficient by using various tools based on language models.

Competence-based learning status measurements: In addition to the knowledge gaps that are already recognized by the bettermarks system, a BKT model (Bayesian Knowledge Tracing) makes it possible to measure the level at which a particular skill is mastered by a student. Based on this information, targeted tasks can be selected that support the student without under- or overtaxing them. Teachers can also use skills-based reporting to gain a deeper understanding of their class and tailor their lessons accordingly.

Metadata enrichment: Metadata can be generated based on anonymized processing data of the exercises, such as the contextualized solution probability, complexity or the procedural competence on which a task is based. Various traditional and AI-based methods have been developed to automatically collect this data. The insights gained in this way are extremely valuable for both product development and editing.

Christophe Speroni
CMO / CIO,
bettermarks GmbH

Titanom supports bettermarks in the development of a revolutionary adaptive textbook that uses the latest methods of machine learning / natural language processing. Titanom’s employees are characterized by their combination of outstanding technical expertise and user-centered product development methodology. This makes them a perfect fit for our philosophy. This has enabled us to better utilize the potential of the technologies and achieve impressive results even faster.

Alexander Berkov
Web Developer,
Titanom Technologies

In working with bettermarks, I am particularly pleased that as a young team member I can have an open and honest exchange about new technologies and ideas with the experienced team at bettermarks. By working closely with the product managers and UX designers, I can make a contribution to the product that goes beyond my pull requests. What also impresses me about bettermarks is how well thought-out the content creation process is, so that content authors can develop varied, digital tasks with adaptive validation logic without extensive prior technological knowledge. I am very grateful to be able to contribute to the creation of innovative educational media in this productive environment.

WOULD YOU LIKE TO FIND OUT MORE? PLEASE CONTACT US!

We look forward to getting to know you. In an initial meeting, we will check whether the missions of our companies have a common core. For us, this is the basis of a successful partnership.

LUKA VUKSANOVIC

[email protected]

Feel free to contact me!

We look forward to getting to know you. In an initial meeting, we will check whether the missions of our companies have a common core. For us, this is the basis of a successful partnership.