Artificial Intelligence in Learning Environments

Digital learning systems are becoming increasingly popular in academic environments. Several different types of technologies and platforms offer diverse approaches in developing learning materials, facilitate learning processes and enhance human resource teams. Adaptability and personalization in Learning Systems become more and more crucial and can be integrated using Artificial Intelligence (AI) technologies. Fraunhofer FOKUS’ business unit, FAME, offers AI-driven solutions for any type of Learning System and its components. 

Artificial Intelligence in Adaptive Learning Environments

Fraunhofer FOKUS researches and develops learning technologies that satisfy the needs of learners, pedagogical staff, content creators, managers of educational institutions and external stakeholders, such as employers and personnel developers. 

Learners can then have adaptive access to interactive course materials and relevant services on various devices (e.g. mobile devices, computers and even Smart TVs) via a seemless frontend.

AI is one of the key drivers of adaptive learning. Analyzing usage data leads to a better understanding of the learning process and thus, optimized content, teaching and learning thereafter. For instance, in digital learning environments, real-time analyses of students’ interactions with learning objects provides vital information about content validity, teaching methods and personal learning habits. All user interactions (for example, with learning objects) are persisted as ADL xAPI or IMS CALIPER statements in a learning record store. To address the evolution of sustainable software architectures, FOKUS researchers developed the Common Learning Middleware (CLM), a field-proven enabler for interoperable microservice infrastructures. This service mediator is strongly recommended in order to manage AI services, data storage, content databases and adaptive user interfaces. 


Related Links:
fame, solutions, artificial intelligence, ai, learning environments, learning technologies
User interface shows Learning Analytics Fraunhofer FOKUS

Learning Analytics

Based on the latest research on learning behaviors, our Learning Analytics algorithms can help both learners and educators to keep track of the learning progress.

For example, our researchers apply unsupervised machine learning techniques, such as X-Means and K-Means clustering-based algorithms, to gather xAPI learning records in order to identify typical learning behaviors and students at risk. The Learning Analytics component can also calculate and visualize the learner’s needs and typical studying behavior. This way, learners can keep track of their personal predicted knowledge level for different learning objects at any point in time. Likewise, teachers can use this data to get an overview of the learners’ overall progress and be aware of any potential knowledge gaps. Content providers, in turn, analyze this data to optimize existing content and obtain recommendations on success criteria for the development of new digital learning materials.  

Recommender Systems

The analyses of learning records, user profiles and content metadata set the basis for Learning Recommender Systems.

Thereby, the Recommender Systems follow a particular paradigm that is different from e-commerce or entertainment domains. These particular systems must consider the unique characteristics that educational settings comprise of, such as the learners’ motivation and needs, typical user behaviors and resulting activity data. As such, Learning Recommender Systems should enable more efficient and effective learning for its users and in turn, future learning activities in a positive manner. In other words, rather than being based on certain preferences, such systems should be based on the actual user’s learning needs.

To meet the unique requirements in the specialized paradigm, our Smart Learning Recommender System utilizes a novel knowledge-based filtering approach that combines the strengths of content-based and collaborative filtering techniques. The algorithms transfer multi-contextual activity data into time-dependent user models. The analyzed contexts may include interactions with and processing times of learning media, self-assessments, exercise performances, fullfiled knowledge prerequisites, lecture times and exam relevance. Additionally, more advanced models may consider the learning patterns of other learners as well as predict the individual forgetting effect after each learning iteration. The resulting scores represent the individual user’s need to learn specific learning objects at a particular point in time of the course. The recommendations are presented to the user as hints for learning that are well explained ( explainable AI) and easy to understand. In comparison to the evaluation results of other educational recommender systems (built with the same datasets), our knowledge-based filtering approach generates the most accurate and time-sensitive recommendations. 

fame, solutions, artificial intelligence, ai, learning environments, learning technologies
Visualization of Learning Paths  Fraunhofer FOKUS

Personalized Learning Paths

Recommender Systems have the ability to suggest upcoming learning objects, but do not consider the sequence of learning. Personalized Learning Paths fill this gap by providing the user with a choice of appropriate learning sequences. 

Fraunhofer FOKUS’ researchers designed a generator to provide personalized learning paths through knowledge networks. These paths are tailored for each learner by taking into consideration the given didactical sequences, pre-acquired knowledge, fulfilled prerequisites, content analyses of similar content, as well as learning activities of other successful learners. Thus, personalized paths are generated at the time at which they are requested, including suggested alternate routes in order to provide the user with a selection of preferred learning items. 

fame, solutions, artificial intelligence, ai, learning environments, learning technologies
Learning Evaluation Fraunhofer FOKUS

Evaluation of Algorithms for Learning

Learners and learning activities frequently change over time, therefore, Learning Recommender Systems need be adaptable in calculating optimal recommendations. Research has found that the evaluation and comparison of learning recommendations lack standardized definitions of datasets and evaluation procedures.

While the standard cross-validation setting typically produces reliable results, it does not work within an educational setting, where, for instance, every topic is relevant (especially for final exams). Furthermore, analyses of closed-course settings indicate a high time dependency. This is where our Time-Dependent Learning Evaluations Framework comes into play, which allows for a reliable and comparable evaluation of the algorithms and a better understanding of temporal aspects in the data. For example, the Timeliness Deviation measure can investigate the time gap between when a learning object is recommended and when it is actually needed by the user. 

fame, solutions, artificial intelligence, ai, learning environments, learning technologies
Chat with LMS Chatbot Fraunhofer FOKUS

Chatbots as Virtual Mentors

Another essential approach in utilizing artificial intelligence methods is to offer virtual learning assistance in form of chatbots, with which users can communicate in textual natural languages. 

Chatbots can help to reduce barriers and answer frequently asked questions more efficiently. They are predominantly based on the Natural Language Understanding (NLU) and Natural Language Generation (NLG) methods, both of which are subsets of Natural Language Processing (NLP). By using Chatbots, learners do not have to log into a learning platform. Instead, they can access their learning content, definitions and content recommendations directly through a text messenger. In addition to answering common organizational and subject matter questions, chatbots can also access other AI services, such as learning analytics, learning recommendation systems, or learning path generators via the middleware. This way, users receive direct feedback on their learning behavior, recommendations for upcoming topics, and general learning reminders (via push notifications).

Key Facts

  • AI components are crucial for adaptive and personalized learning environments.
  • Learning Analytics algorithms help both learners and educators to make learning more efficient and effective.
  • The Smart Learning Recommender uses a novel knowledge-based filtering approach that generates exceptionally accurate and time-sensitive recommendations.
  • The Time-Dependent Learning Evaluations Framework allows for a better understanding of temporal aspects, thus making Learning Recommender Systems more adaptive to individual learning needs.
  • In addition to recommending learning tasks, Personalized Learning Paths suggest appropriate learning sequences as well.
  • Chatbots relieve educators in efficiently answering frequently asked questions and represent an easy accessible way to reach other AI services.
  • The Common Learning Middleware (CLM) manages the data flow, AI services, data storage, content databases and adaptive user interfaces.

If you have any further questions or interest in integrating our AI-powered learning technologies in your environment, feel free to contact us! All components are available for licensing and our experts at Fraunhofer FOKUS also offer the design and development of new concepts and prototypes. Furthermore, we are happy to share our expertise (i.e. workshops) and provide consulting and research support for feasibility studies.