My name is In Son Zeng, I come from Macau, and now I am a researcher working in the field of AI in Education at the University of Michigan. Before working in the educational field, I graduated from the University of Michigan with a Master of Science degree in Applied Statistics and from the University of Macau with a bachelor's degree in Mathematics.
Since 2015, I have had opportunities to teach many data science students who are full with research goals to transform the future in a variety of settings: K-12 education, academia, healthcare, automotive industry, environmental protection, cognitive science, and social science. Having instructed 200+ students in several Python programming, mathematics and statistics courses at the University of Michigan as a graduate student instructor, students graciously told me about how they approach the challenge of disseminating data-driven knowledge publicly at low cost and with high interactivity.
Data science now encompasses a superfluous variety of technologies and techniques. They are, but not limited to, visualization, data manipulation, modeling, version control, storytelling, data elicitation and system administration. The advancement of computation technology has allowed:
As my research goal lies in personalized learning at scale, I am incessantly exploring effective and efficient ways of augmenting data science education at scale through human intelligence, adaptive pedagogy and literate programming approaches. In the data manipulation course, I assigned students into small, roundtable groups of 5 people. Each group would be responsible for identifying a topic and problems of interest, curating data, writing a program to analyze/visualize and sharing findings to the class with interactive questioning-answering time. In collaborative group learning, students could quickly identify data science workflow, recognize the ways of solving problems through teamwork and evade misconceptions and pitfalls on the program through immediate feedback from teammates (otherwise computer software might not remind students). In short, collaboration proved to be effective in enhancing data science learning experience through collaboration. In addition, I believe that motivation plays a pivotal role in acclimating students to data science practices. In the previous courses I taught, I encouraged critical thinking about the authenticity of data science and literacy in statistical interpretation. Most students could pinpoint data-related problems within a short amount of time upon importing data locally, or scraping the data from websites if there are. Each time I visit a topic in a class, I will always address questions including:
My educational passion also involves designing multi-modality MOOCs (Massive Open Online Course) to disseminate preeminent concepts, methodologies and technical training with low cost. I started to develop the Applied Probabilistic Data Science in Python specialization in January 2021 in Bayesian Statistics and Data Visualization. Targeting to novice programming learners and industry professionals who have less than 1 year of data science experience, I tailored widely-known Bayesian thinking, estimation techniques and algorithm-based experiments in a content-heavy and math-light fashion. Since the enrollment of online learning students from outside the U.S. has more than doubled during the last 5 years, I localized teaching examples in video lectures to more than 10 distinct countries, and shaped the course activities in a participatory approach. This created opportunities for learners to co-create knowledge and gain ownership of excellent examples that they offer from their funds of knowledge to other learners. I dedicate myself to exploring how sustainable learning communities can be created at the intersection of the areas of culturally relevant content, psychological sense of ownership, and learning analytics in my very first MOOC in Coursera (currently undertaking beta testing).
My educational vision “Make the world see!” and design principle “Start small, grow big” responds to the widening educational inequality due to pandemic and wealth gap. Starting from Bayesian statistics, I am planning to offer courses such as Data Manipulation, Machine Learning, Text Mining, Information Retrieval, Machine-Assisted Instructional Design, and Artificial-Intelligence-Assisted Educational Technology to grow the community of university students with the awareness of the rising real-life technologies. To extend the programming education context to a wider public, I am actively collaborating with digital writers, data scientists, researchers, and educational researchers to transform traditional coding concepts and theories into humorous notebooks, interactive animations, comics and slides that reduce the cognitive load for students to fathom the instructional contents.
While AI technologies are profoundly impacting society around the globe, the applications of artificial intelligence are not immune to the risks of pitfalls of data-driven and AI-supported tools. I’m aware of and at the critical time to respond to academic questions through educational practice such as 1) How can we prepare K-12 students for an AI-permeated future? 2) How to enhance algorithmic fairness when predicting a student's learning outcome and behaviors? 3) How to arouse deeper dialogues among researchers working on AI-supported education and educational technology of the “Fairness, Accountability, Transparency and Ethics”, the FATE of AI around the ethical and societal implications of AI? The Educational Technology Collectives lab I worked with deeply engages with the concerns about the present and future roles of AI in education. I hope to dedicate as a lecturer for Data Science and Artificial Intelligence related courses in the upcoming years to fostering the core values of fairness, authenticity, transparency and equity in data science and artificial intelligence.