PhD student work at FDG and CIG conferences

The work of IDG Ph.D. student Daniel Karavolos has been presented in two conferences during August 2018: "Using a Surrogate Model of Gameplay for Automated Level Design" was presented in the international conference on Computational Intelligence in Games, in Maastricht, NL, while "Pairing Character Classes in a Deathmatch Shooter Game via a Deep-Learning Surrogate Model" was presented at the FDG workshop on Procedural Content Generation, in Malmö, Sweden. Both papers were co-authored with IDG staff members Antonios Liapis and Georgios N. Yannakakis, and propose using a computational model that can predict a game's outcome by observing the level and character classes instead of simulating gameplay with AI agents. The deep learning model is shown to work as a surrogate to simulations when attempting to automatically adapt a level design to specific competing classes, as well as when generating the appropriate classes for a specific level.