Procedural Content Generation of Puzzle Games using Conditional Generative Adversarial Networks

Abstract

In this article, we present an experimental approach to using parameterized Generative Adversarial Networks (GANs) to produce levels for the puzzle game Lily's Garden 1 . We extract two condition-vectors from the real levels in an effort to control the details of the GAN's outputs. While the GANs performs well in approximating the first condition (map-shape), they struggle to approximate the second condition (piece distribution). We hypothesize that this might be improved by trying out alternative architectures for both the Generator and Discriminator of the GANs.

Adversarial Random Forest Classifier for Automated Game Design

“…Initially PCGML research focused on level generation rather than rule generation or autonomous game design. In our system, we take inspiration from generative adversarial networks (GANs), which have been used in level generation [8], for our adversarial training methodology. We lack the space to discuss GANs in detail.…”

Illuminating Mario Scenes in the Latent Space of a Generative Adversarial Network

“…Instead, our focus is on assessing the performance of QD algorithms in generating a variety of scenes with desired characteristics, and in measuring modern MAP-Elites variants that excel at the exploration of continuous domains. Our work is also related with conditional generative models (Hald et al 2020;Snodgrass and Ontañón 2014;Ping and Dingli 2020). While it is possible to condition GANs on desired BCs, there is no guarantee that the generated scenes will have the properties specified by the conditioning input.…”