Search for: All records

MineObserver 2.0 is an AI framework that uses Computer Vision and Natural Language Processing for assessing the accuracy of learner-generated descriptions of Minecraft images that include some scientifically relevant content. The system automatically assesses the accuracy of participant observations, written in natural language, made during science learning activities that take place in Minecraft. We demonstrate our system working in real-time and describe a teacher dashboard to showcase observations, both of which advance our previous work. We present the results of a study showing that MineObserver 2.0 improves over its predecessor both in perceived accuracy of the system's generated descriptions as well as in usefulness of the system's feedback. In future work, we intend improve system generated descriptions to give more teacher control and shift the system to perform continuous learning to more rapidly respond to novel observations made by learners. 

Our work investigates interest triggering, a necessary component of sustaining and developing long-term interest in STEM. We gathered interview data from middle school aged learners (N = 7) at a science-focused Minecraft summer camp over a period of one week. We first identified STEM interest triggering episodes, then categorized each episode based on codes developed previously by Renninger and Bachrach (2016). Our initial findings show differences in the frequency of interest triggering episodes across individuals and suggest that personal relevance and the use of Minecraft played prominent roles. 

The importance of promoting interest in STEM is crucial to the recruitment and retainment of underrepresented populations in the STEM field. We created a one-week summer camp program centered around astronomy using Minecraft to help promote interest in science. We capitalize on data logs collected from two 1-week camps in summer 2019 and code scientific observations made by campers for the types of scientific observations and their level of quality using Cohen’s Kappa. Results showed that the majority of observations are descriptive, comparative, inferential, or analogous, as opposed to being off topic or factual. We discuss possible reasons for this distribution and design implications for future reiterations.