Crowdsourcing Community Intelligence with Interactive Streaming

Distributed Computing / Human-Guided AI

The Game: Through mini-games and streaming communities, we crowdsourced human intelligence to gather crucial information that helps in processing data for multi-drug resistant cancers and macular degeneration. With a combination of NodeJS and a Minecraft Spigot Server we empowered crowd intelligence in MIXER-Interactive elements through a custom Minecraft Mini-Game. For the first time ever, applying these technologies with a unique approach allowed us to engage streamers with their viewers in purposeful play.
The Science: This Human Computational Gaming technique was built into a twin stick shooter-style game and a Minecraft mini-game, both of which performed cluster analysis on chemotherapeutic co-medication data. The clustering games demonstrate the platform’s problem-agnostic abilities by focusing on data standards and algorithm interfaces. Additionally, we created a collaborative machine learning computer vision technique that allowed humans to apply a perception filter to a learning algorithm to increase the accuracy of automated analysis for optical coherency tomography (OCT) scans used in macular degeneration diagnosis and treatment.