Undergraduate Research Assistant Drew Tumblin

Undergraduate Research Assistant Drew Tumblin

Photo of Drew Tumblin

Undergraduate Research Assistant Drew Tumblin

Drew's current projects include AI Beatmap Generation for Rhythm Games. This project aims to develop an AI-driven system capable of generating beatmaps that are virtually indistinguishable from those created by humans. The current implementation leverages a Convolutional Neural Network (CNN), and achieved a nearly 50% pass rate in tests with human subjects. Moving forward, the primary objectives are to enhance the AI by transitioning to a modern Generative Adversarial Network (GAN). This update is expected to improve the quality and realism of the generated beatmaps. Additionally, expanding the model's compatibility to include other rhythm games is a key priority, broadening the application's utility and impact in the rhythm game community.

Interests: Artificial Neural Networks, Rhythm detection and beat placement, and Generative Adversarial Networks.

See also: Undergraduate