USC GamePipe Laboratory Seminar Series

Title: Automated Game Generation via Machine Learning
Wednesday, 24 October 2018, 6:00pm
Speaker: Matthew Guzdial, Georgia Tech
Location: USC GamePipe Laboratory, EGG-108

Abstract
Automated game design has remained an open question in the field of Game AI. Prior automated game design approaches have relied on hand-authored or crowd-sourced knowledge, which limits scope and application. Our approach instead relies on machine learning to derive approximate representations of games, in terms of level design and mechanics, from gameplay video. Our approach recombines knowledge from these learned representations to create new games via conceptual expansion, a novel algorithm inspired by human creativity.

Bio

Matthew Guzdial is a Computer Science PhD candidate in the School of Interactive Computing within the College of Computing at Georgia Tech. His research focuses on the intersection of machine learning and creativity, with the majority of his thesis work devoted to the automated generation of games. He is the recipient of a Unity Graduate Fellowship and invitee to the 2018 Heidelberg Laureate Forum. His work has been featured in BBC, WIRED, and Popular Science.