Designing Adaptive Argumentation Feedback Systems based on Machine Learning
The EDIC program is happy to invite you to a public talk by our post-doctoral student Thiemo Wambsganss at the Machine Learning for Education Lab at EPFL.
The aim of the talk is to present his achievements to a broad audience to prepare for hiring interviews coming up soon.You are warml welcome to listen to the talk and participate in the Q&A session at the end of the presentation.
Argumentation is an omnipresent rudiment of daily communication and thinking. The ability to form convincing arguments is not only fundamental to persuading an audience of novel ideas but also plays a major role in strategic decision-making, negotiation, and constructive, civil discourse. However, humans often struggle to develop argumentation skills owing to a lack of individual and instant feedback in their learning process, since providing feedback on the individual argumentation skills of learners is time-consuming and not scalable if conducted manually by educators. To investigate if dynamic technology-mediated argumentation feedback improves persuasive writing, we built a novel dynamic argumentation feedback system based on machine learning (ML). The novel system provides learners with dynamic writing feedback opportunities based on logical argumentation errors irrespective of instructor, time, and location. We evaluate our system in two studies to test if dynamic argumentation feedback improves persuasive writing performance more so than traditional upfront argumentation instruction (H1) and if dynamic feedback on repeated argumentation tasks (over three months) leads to better learning in comparison to static feedback (H2). Our results show that dynamic feedback helps learners to increase their metacognitive argumentation skills across domains compared to the benchmark of upfront instructions and static feedback. Our research, thus, investigates the potential of dynamic feedback tools in a field study to support students to train their skills in large-scale or distance-learning scenarios. This work can support researchers in designing new skill learning systems based on ML to leverage these systems’ full potential not only for information systems but also for metacognition skill-based future (continuous) education.
Thiemo Wambsganss is a PostDoc at the Machine Learning for Education Lab at EPFL, advised by Prof. Dr. Tanja Käser. His research interests lie at the intersection of Natural Language Processing (NLP), Human-Computer Interaction (HCI), and Educational Technology. Here, he is primarily driven by the vast opportunities to enhance and improve pedagogical scenarios based on recent advantages in NLP and
Machine Learning to enable students to learn when, where, and how they want independent of an educator or their background. To do so, he uses techniques from Artificial Intelligence such as Transfer and Deep Learning to build AI-powered education tools such as Intelligent-Tutoring-Systems, Conversational Agents, and intelligent writing support systems. His publications in the area of Argumentation Writing Support, Empathy Detection, and Pedagogical Conversational Agents are mainly in the areas of HCI (e.g., CHI20, CHI21, CHI22), NLP (ACL21, ACL22, COLING22), and Information System (ICS20, ICIS21, ICIS22) and have received several awards, such as the Delina Learntec Award 2021, the Best Theory Paper First Runner-Up Award at ICIS20 or two ACM CHI Honorable Mention Awards. Thiemo completed his PhD at the Institute of Information System at the University of St.Gallen.