Game generation and analysis has commonly relied on hand authored rules and heuristics. This authoring task comes with a high authorial burden, both in the amount of rules and heuristics that need to be authored for decent coverage and in the complexity of authoring these rules. In this paper I present early work on \textit{Leda} and inductive logic programming system designed to learn these rules, so as to support further generation and analysis. I present Leda, describe its process, and finally show a sample set of the rules that it learns.

Markov chains are an enticing option for machine learned generation of platformer levels, but offer poor control for designers and are likely to produce unplayable levels. In this paper we present a method for guiding Markov chain generation using Monte Carlo Tree Search that we call Markov Chain Monte Carlo Tree Search (MCMCTS). We demonstrate an example use for this technique by creating levels trained on a corpus of levels from Super Mario Bros. We then present a player modeling study that was run with the hopes of using the data to better inform the generation of levels in future work.

Procedural Content Generation (PCG) using machine learning is a fast growing area of research. Action Role Playing Game (ARPG) levels represent an interesting challenge for PCG due to their multi-tiered structure and nonlinearity. Previous work has used Bayes Nets (BN) to learn properties of the topological structure of levels from The Legend of Zelda. In this paper we describe a method for sampling these learned distributions to generate valid, playable level topologies. We carry this deeper and learn a sampleable representation of the individual rooms using Principal Component Analysis. We combine the two techniques and present a multi-scale machine learned technique for procedurally generating ARPG levels from a corpus of levels from The Legend of Zelda.

A touted use of Procedural Content Generation is generating content tailored to specific players. Previous work has relied on human identification of player profile features which are then mapped to level generator features. We present a machine-learned technique to train generators on Super Mario Bros. videos, generating levels based on latent play styles learned from the video. We evaluate the generators in comparison to the original levels and a machine-learned generator trained using simulated players.

Procedural Content Generation (PCG) has seen heavy focus on the generation of levels for video games, aesthetic content, and on rule creation, but has seen little use in other domains. Recently, the ready availability of Long Short Term Memory Recurrent Neural Networks (LSTM RNNs) has seen a rise in text based procedural generation, including card designs for Collectible Card Games (CCGs) like Hearthstone or Magic: The Gathering. In this work we present a mixed-initiative design tool, Mystical Tutor, that allows a user to type in a partial specification for a card and receive a full card design. This is achieved by using sequence-to-sequence learning as a denoising sequence autoencoder, allowing Mystical Tutor to learn how to translate from partial specifications to full.

Procedural Content Generation (PCG) has been a part of video games for the majority of their existence and have been an area of active research over the past decade. How- ever, despite the interest in PCG there is no commonly ac- cepted methodology for assessing and analyzing a generator. Furthermore, the recent trend towards machine learned PCG techniques commonly state the goal of learning the design within the original content, but there has been little assess- ment of whether these techniques actually achieve this goal. This paper presents a number of techniques for the assess- ment and analysis of PCG systems, allowing practitioners and researchers better insight into the strengths and weaknesses of these systems, allowing for better comparison of systems, and reducing the reliance on ad-hoc, cherry-picking-prone tech- niques.

Procedural content generation via machine learning (PCGML) has shown success at producing new video game content with machine learning. However, the majority of the work has focused on the production of static game content, including game levels and visual elements. There has been much less work on dynamic game content, such as game mechanics. One reason for this is the lack of a consistent representation for dynamic game content, which is key for a number of statistical machine learning approaches. We present an autoencoder for deriving what we call "entity embeddings", a consistent way to represent different dynamic entities across multiple games in the same representation. In this paper we introduce the learned representation, along with some evidence towards its quality and future utility.

Techniques for procedural content generation via machine learning (PCGML) have been shown to be useful for generating novel game content. While used primarily for producing new content in the style of the game domain used for training, recent works have increasingly started to explore methods for discovering and generating content in novel domains via techniques such as level blending and domain transfer. In this paper, we build on these works and introduce a new PCGML approach for producing novel game content spanning multiple domains. We use a new affordance and path vocabulary to encode data from six different platformer games and train variational autoencoders on this data, enabling us to capture the latent level space spanning all the domains and generate new content with varying proportions of the different domains.

In this work, we present TOAD-GAN (Token-based One-shot Arbitrary Dimension Generative Adversarial Network), a novel Procedural Content Generation (PCG) algorithm that generates token-based video game levels. TOAD-GAN follows the SinGAN architecture and can be trained using only one example. We demonstrate its application for Super Mario Bros. levels and are able to generate new levels of similar style in arbitrary sizes. We achieve state-of-the-art results in modeling the patterns of the training level and provide a comparison with different baselines under several metrics. Additionally, we present an extension of the method that allows the user to control the generation process of certain token structures to ensure a coherent global level layout. We provide this tool to the community to spur further research by publishing our source code.

It's been 10 years since the first ever Mario AI Competition, so I return to the world of Super Mario level generation research and catch up one some of the more interesting examples that have arisen in recent years. This video is inspired by the following AI research papers and projects: NOOR SHAKER: http://lynura.com/publications.php It's is supported through and wouldn't be possible wthout the wonderful people who support it on Patreon. You can follow AI and Games (and me) on Facebook, Twitter and Instagram: http://www.facebook.com/AIandGames