Chair: Prof. Nancy Law, Professor, Faculty of Education, The University of Hong Kong
I will present our research and open source data analytic tools regarding how students learn, especially computational thinking and programming, on Massive Open Online Courses (MOOC) and a blended K-12 course. We use machine learning and data science to study this. In one study, examining how student learning is influenced by active practice, e.g. “doing” exercises, as compared to passively learning by video watching. We also consider scenarios such as how learning designs, prior experience, duration, and course content impact learning behavior and, in turn, how this behavior correlates with an assessment grade. Moreover, we have developed an open source framework for data science for MOOCs, MOOC-Learner-Project (http://mooclearnerproject.csail.mit.edu/). This has components that transform “raw” edX platform clickstream and other behavioral data streams into tables organized according to the MOOCdb data model, support feature engineering, visualization and modeling via machine learning.
Dr. Erik Hemberg is a research scientist in the AnyScale Learning For All (ALFA) group at Massachusetts Institute of Technology Computer Science and Artificial Intelligence Laboratory. He has a PhD in Computer Science from University College Dublin, Ireland and a MSc from Chalmers University of Technology, Sweden. At ALFA he focuses on scalable machine learning, artificial intelligence and data science.
The ALFA group has projects in MOOC technology, cyber security and clinical medicine knowledge discovery. He developed a blended Learning course at Shantou University and computer science outreach course at University College Dublin. His paper regarding the Doer-effect received an award at the LWMOOC 2019 conference