At the intersection of computer science and cognitive science, we conduct applied research into human-centered explainable AI and intelligent data-intensive systems and their application in technology-enhanced learning and knowledge management domains. Headed by Prof. Dr. Mohamed Amine Chatti, we take an interdisciplinary approach in designing and developing trustworthy human-centered AI technologies and novel analytics-driven methods and visualizations for tomorrow’s smart learning and working environments. We offer lectures, seminars, and practical courses in learning technologies, web technologies, data science, visual analytics, and human-AI interaction following a technology-enhanced, student-centered learning approach.

We're part of the Department of Human-centered Computing and Cognitive Science (HCCS), Faculty of Computer Science at the University of Duisburg-Essen, one of the top universities in Germany.

News & Events

The Social Computing Group was present at the 32nd ACM Conference on User Modeling, Adaptation and Personalization (UMAP'24) in Cagliari, Sardinia, Italy, from July 1-4, 2024. UMAP is the premier international conference for researchers and practitioners working on systems that adapt to individual users, to groups of users, and that collect, represent, and model user information. We presented our ongoing work on GNN-based user modeling in the context of our project CourseMapper. At the 1st International Workshop on User-Centered Practices of Knowledge Discovery in Educational Data (UKDE 2024), Rawaa Alatrash presented our paper titled "Transparent Learner Knowledge State Modeling using Personal Knowledge Graphs and Graph Neural Networks", available in UMAP24 Adjunct Proceedings here.

On June 19, 2024, Shoeb Joarder attended the event Learning Analytics in Practice (LAP24). LAP24 is a 24-hour, global online event that is intended to highlight and feature learning analytics practitioners around the world discussing and showcasing learning analytics in practice. Shoeb presented our ongoing research on Human-Centered Learning Analytics (HCLA) and gave a demo on the human-centered design and implementation of learning analytics indicators using OpenLAP. We received positive feedback from LAP24 participants, demonstrating the potential of the OpenLAP ISC Creator and OpenLAP Editor to actively involve different stakeholders in the learning analytics process. A demo can be found here.

Our paper "Transparent Learner Knowledge State Modeling using Personal Knowledge Graphs and Graph Neural Networks" got accepted at the workshop on User-Centered Practices of Knowledge Discovery in Educational Data (UKDE 2024) in conjunction with the 32nd ACM Conference on User Modeling, Adaptation and Personalization (ACM UMAP 2024). A preprint version of the paper is available here.

Just in time for the start of the summer semester of 2024, we are super excited to announce the launch of the pilot phase of our MOOC platform CourseMapper. CourseMapper will be used this semester to support collaborative learning in our Advanced Web Technologies (AWT) lecture course and Interactive Data Exploration and Analytics (IDEA) lab course.

Please feel free to test CourseMapper yourself here. If you have any questions, need more information, would like to join us in the further development of CourseMapper, or would like to collaborate on a research project around CourseMapper to democratize learning, please do not hesitate to contact us.

The Social Computing Group was present at the prestigious International Learning Analytics and Knowledge Conference (LAK24) in Kyoto, Japan, from March 18-22, 2024. LAK is the premier conference in the field of learning analytics and educational data science. We presented two papers in the context of our ongoing projects CourseMapper and OpenLAP. Qurat Ul Ain presented our full paper "Learner Modeling and Recommendation of Learning Resources using Personal Knowledge Graphs", available as open access here. Shoeb Joarder presented our poster paper "A No-Code Environment for Implementing Human-Centered Learning Analytics Indicators", available in LAK24 Companion Proceedings here.

In our new paper "ConceptGCN: Knowledge concept recommendation in MOOCs based on knowledge graph convolutional networks and SBERT", recently published in Computers and Education: Artificial Intelligence, we propose ConceptGCN, an end-to-end framework that combines knowledge graphs, graph convolutional networks, and transformer sentence encoders to enhance the representations of knowledge concepts and learner models and provide personalized and explainable recommendation of concepts in our MOOC platform CourseMapper. The paper is available as open access here.

On December 1, 2023, Shoeb Joarder attended the Lehrkräftefortbildung Informatik 2023. He presented our ongoing work on Human-Centered Learning Analytics (HCLA) with Open Learning Analytics Platform (OpenLAP), our platform for self-service learning analytics. He showed a demo of OpenLAP in action, with a focus on the Indicator Specification Card (ISC) Creator and the Indicator Editor in OpenLAP.

Our paper "Interactive Explanation with Varying Level of Details in an Explainable Scientific Literature Recommender System" has been published in the International Journal of Human-Computer Interaction (IJHCI). In this work, we show how to design explanations with different levels of detail, in a systematic and theoretically-sound manner. Further, we highlight the positive impact of user-controlled, interactive explanation with varying level of details on the perception of explainable recommendation. A preprint version of the paper is available here.

Our paper "Automatic Construction of Educational Knowledge Graphs: A Word Embedding-Based Approach" has been published in the Information Journal.  In this work, we aimed at improving the concept extraction and weighting mechanisms in educational knowledge graphs (EduKGs) by leveraging state-of-the-art word and sentence embedding techniques. The paper is available as open access here.

Our paper "Semantic Interest Modeling and Content-Based Scientific Publication Recommendation Using Word Embeddings and Sentence Encoders" has been published in the Multimodal Technologies and Interaction Journal. In this work, we combine word embedding-based keyphrase extraction techniques with knowledge bases to generate semantically-enriched user interest models, and we leverage pretrained transformer sentence encoders to represent user models and papers and compute their similarities.  The paper is available as open access here.  

On August 28 & 29, 2023, Shoeb Joarder attended the 2nd Learning AID at the Ruhr-University of Bochum. The conference and workshop was organized by the state funded project KI:edu.nrw – a project to promote Didactics, Ethics and Technology of Learning Analytics and AI in Higher Education.  The event focused on the practical use of learning analytics, artificial intelligence, and data mining in higher education. Shoeb presented our ongoing research on Human-Centered Learning Analytics (HCLA) and gave a demo on the ISC Creator module in our Open Learning Analytics Platform (OpenLAP). The ISC Creator provides an intuitive user interface that allows the low-cost design of low-fidelity Learning Analytics indicators using Indicator Specification Cards. A demo can be found here.

You might be interested in understanding the difference between Transparency and Justification in explainable recommender systems. Check out our contribution "Justification vs. Transparency: Why and How Visual Explanations in a Scientific Literature Recommender System" which is published in the Information Journal  as part of the Special Issue "Information Visualization Theory and Applications". The paper is available as open access here.

Earlier News & Events