AI in the education sector: A project with EduPlex
This project demonstrates how artificial intelligence (AI) can be used in the field of education and training to personalize the learning experience. Specifically, the goal was to improve an existing learning platform by recommending courses to learners that match their needs and preferences.
About the stakeholders
EduPlex is a research initiative in the field of education. For this project, EduPlex brought together a diverse consortium of companies and educational institutions.
Courseticket is an online course portal that builds digital L&D platforms and apps. The company provides the technical foundation for well-known educational providers, publishers, SMEs, corporates, universities, and ministries.
Hochschule Schmalkhalden has a strong focus on artificial intelligence (AI) and offers a variety of programs and research initiatives in this area. The university emphasizes the combination of theoretical knowledge and practical application to optimally prepare students for their professional future.
WBS Training AG is a German educational provider based in Berlin, specializing in adult education and training, with approximately 250 branches in Germany.
The team from ONTEC AI accompanied the consortium as a subcontractor in an advisory position. The team, consisting of AI experts, develops, accompanies, and advises medium and large companies on the implementation of secure AI systems.
Challenge and goal
The market for education, training, and further education is highly competitive in times of digital access to learning materials. Large providers like YouTube Learn face smaller learning platforms, which must particularly distinguish themselves to be sustainably successful.
The learning process experience is crucial for success.
Courseticket faced exactly this challenge.
The existing system could not recommend courses tailored to individual goals and preferences, leading to lower user engagement and suboptimal learning outcomes.
Therefore, the existing learning platform needed to be optimized to offer its users a particularly personalized and engaging experience.
The project’s goal was to develop an AI-based recommendation system that efficiently suggests relevant courses to improve user satisfaction and support professional development through retraining and further education measures.
The AI-driven recommendation system would automatically suggest suitable courses to learners, support self-directed learning, and promote the use of small learning units, which are crucial for the future of digital learning environments. This personalization aims to increase motivation through tailored learning paths.
The ability of AI to process novel inputs, improve with more data, and automate tasks increases speed and efficiency, enabling companies to scale quickly.
For the organization, implementing this system would increase overall revenue and improve capabilities in AI-based recommendation systems. It should also help attract top talent.
Project process and methods
For the project, the different stakeholders in the consortium came together to contribute their expertise.
ONTEC AI’s contribution consisted of providing advice on developing an AI system that recommends suitable courses based on user data. Specifically, the AI consulting part included 6 workshops in which various stakeholders discussed their expectations, requirements, and recommendations for AI development.
Methodologically, a mix of TOGAF (The Open Group Architecture Framework, a methodology for IT enterprise architecture), DISR (Design-oriented Information System Research, a research method by Österle et al.), and RFISR (Research Framework for Information System Research, a research method by Nunamaker & Chen) was used to define the following work steps:
- Business: Here, the business requirements of the stakeholders for the AI component to be developed were defined.
- Data: Here, the available and usable data sources were identified.
- Objective / Gold Standard: Here, the goals were formally defined, and a target against which development and research would be conducted was set.
- Application Experimentation Loop: An ongoing experimentation cycle to identify the best approaches:
- Select (Blend of) Approaches: Here, it was defined which AI and data science methods would be used for the recommendation.
- Implement: Here, this approach was experimentally implemented.
- Experiment: Here, the approach was measured against the goal and the gold standard. If the goals were not achieved, back to step a.
- Choose Technology: Finally, based on the results, it was identified how exactly the recommendation system would be implemented.
The core of the consulting service was the structuring of tasks according to this methodology and support in developing these steps.
The project was to be carried out within 18 months, with completion set for March 31, 2023.
Result
The result was an AI-based recommendation system that offers a personalized learning experience by suggesting the most suitable courses for users’ goals. Time and cost efficiency and the high quality of recommendations stand out positively.
The recommendation system concept included two main strategies: interaction-based and content-based recommendations.
- Interaction-based recommendations would generate suggestions based on user behavior and use methods such as LDA, basket embeddings, user classification, and sequence learning.
- Content-based recommendations would use actual course content descriptions and apply state-of-the-art techniques such as word embedding with WMD, content embedding with transformers, and cosine-similarity-based search.
Summary and key takeaways
This project demonstrates the transformative potential of AI in the education sector by creating a personalized, efficient, and scalable learning environment for learners and educational institutions.
By exploring the potential of AI in educational technology, this project paves the way for a future where learning is more personal, engaging, and efficient.
- AI enables tailored learning paths that increase learners’ motivation and engagement.
- Automated course suggestions are time and cost-efficient, benefiting both learners and providers.
- By analyzing user behavior and course content, informed decisions can be made that improve the quality and relevance of learning offerings.
- Organizations that integrate AI into their learning platforms improve their market position and can better compete with large competitors.
- AI supports professional development through precise recommendations for retraining and further education measures, leading to better career opportunities.