On the 8th of Feb. 2022 in a well-ventilated seminar room at Leibniz University, the PhD students and Principal Investigators (PIs) of the LernMINT program were finally able to gather in person for 3 days of discussion, poster presentation and collaboration. After many delays and restrictions because of the COVID-19 pandemic, it was great to meet most of our colleagues in person. Among the LernMINT project’s goals is the building of bridges between the research areas of math and science education and that of computer science and machine learning. The value of in person group discussions to facilitate this can not be stressed enough. What follows is a very brief recap of presentations and discussions during the retreat. And some of the lessons we learned.
One of the main content outputs of the retreat were poster presentations on each of the PhD projects. Tai le Quy presented his work on using clustering to partition students into work groups that satisfy fair representation of protected demographic characteristics in each group with a focus on explainability. Evelyn Navarette presented work on supporting informal learning in STEM subjects through educational videos. She aims to automate the finding of effective videos based on features grounded in Multimedia Learning Principles, computer vision and natural language processing. Nico Marten also works on improving informal learning of basic university math with digital media. His work helps students build digital competences to address their knowledge gaps. Also in the field of informal learning and addressing the challenges of first year university students, Lukas M. Günther studies mathematical enculturation in the process of transition from school to university. The project focuses on the analysis of ‘cultural breaks’ experienced by most students, their strategies, and ways of fostering a successful overcoming of these obstacles through affinitive and informal learning processes. Through another theoretical lens, Ludwig Laukert analyzes this shift in terms of the acquisition of structural praxeologies with the goal of building a chatbot to help students attending lectures in analysis with this acquisition.
Evaluation of various kinds and adaptive tutoring constitute a common theme of many of the projects we discussed. Jos Oldag investigates whether and how core chemistry concepts can be identified in student diagrams. Marvin Roski built the learning platform I3learn which which he has aims to take inclusive chemistry teaching in classroom settings to the next level. He uses machine learning to provide students with individualized feedback along their own learning path. In a similar vein, on the Physics didactics side of things, Lukas Dieckhoff aims to develop a digital tutor that would provide individual feedback in the context of a project to investigate the transition from instructed to independent problem solving. Back on the subject of evaluation, Judith Stanja investigates text-based techniques to support formative assessment of students’ statements about adaptation in evolution. Tom Bleckman also investigates automation of the evaluation of formative assessments in the form of drawings, texts, multiple choice questions, etc. in Physics education. Also working on physics education. Vitor Lecio Fontanella models the acquisition and use of terminology in physics teaching using natural language processing (NLP). Also focusing on NLP, Ratan Sebastian works on tools to help K-12 teachers find Open Educational Resources that are relevant to them and their class through curriculum modeling and learning resource segmentation.
One of the most useful outcomes from our discussions was finding overlaps between the various projects. For instance a shared use of NLP in several projects, various approaches to analyze informal learning and a shared need for data management and analysis tools and know-how. Focussed discussions within these smaller groups led to interesting schematisations of various project tasks. In addition to project specific discussions, more general questions about the day to day work of research and the management of the project. One particularly useful presentation that helped bridge the two disciplines represented was one by PIs from each side presenting conventions for paper writing and publishing in Didactics (Physics, Chemistry and Maths individually) and Computer Science. Another useful presentation from the PIs was a lightning round of presentations of their own research which was also valuable to know what kinds of questions could be addressed to whom.
Overall, it was a fruitful and intense 3 days of learning and exchanging ideas. We got a better understanding of all the projects underway in the program, the expertise of the other PhD students and PIs and formed special interest groups for further focussed discussions. It afforded us a vibrant forum to strengthen intra-project research collaborations moving forward.
lernmint.org 