Educating Machine Learning to Game Designers: Accepted CHI 2019 LBW

02.8.2019 |


This fall, I have had the pleasure of working with UX Designer Jiachi Xie on QUBE; an interactive visualizer focusing on machine learning (ML) education for game designer. As a researcher on this project, I am working with Jiachi to explore interactive visualization techniques for ML education. Our recent paper has been accepted into CHI 2019’s Late Breaking Work (LBW). Our paper present our results from two expert panels reviewing QUBE and our next steps for this project. Currently, Jiachi and I are implementing design updates and starting a user study in March!

QUBE Screen Shot QUBE Screen Shot

Interactive Visualizer to Facilitate Game Designers in Understanding Machine Learning

Abstract: Machine learning (ML) is a useful tool to modern game designers but often requires a technical background to understand. This gap of knowledge can intimidate less technical game designers from employing ML techniques to evaluate designs or incorporate ML into game mechanics. Our research aims to bridge this gap by exploring interactive visualizations as a way to introduce ML principles to game designers. We have developed QUBE, an interactive level designer that shifts ML education in the context of game design. We present QUBE’s interactive visualization techniques and evaluation through two expert panels (n=4, n=6) with game design, ML, and user experience experts.