Generative Adversarial Networks (GANs) are a machine learning approach capable of generating novel example outputs across a space of provided training examples. Procedural Content Generation (PCG) of levels for video games could benefit from such models, especially for games where there is a pre-existing corpus of levels to emulate. This paper trains a GAN to generate levels for Super Mario Bros using a level from the Video Game Level Corpus. The approach successfully generates a variety of levels similar to one in the original corpus, but is further improved by application of the Covariance Matrix Adaptation Evolution Strategy (CMA-ES). Specifically, various fitness functions are used to discover levels within the latent space of the GAN that maximize desired properties. Simple static properties are optimized, such as a given distribution of tile types. Additionally, the champion A* agent from the 2009 Mario AI competition is used to assess whether a level is playable, and how many jumping actions are required to beat it. These fitness functions allow for the discovery of levels that exist within the space of examples designed by experts, and also guide the search towards levels that fulfill one or more specified objectives.

Ms. Pac-Man is a challenging video game in which multiple modes of behavior are required: Ms. Pac-Man must escape ghosts when they are threats and catch them when they are edible, in addition to eating all pills in each level. Past approaches to learning behavior in Ms. Pac-Man have treated the game as a single task to be learned using monolithic policy representations. In contrast, this paper uses a framework called Modular Multi-objective NEAT (MM-NEAT) to evolve modular neural networks. Each module defines a separate behavior. The modules are used at different times according to a policy that can be human-designed (i.e. Multitask) or discovered automatically by evolution. The appropriate number of modules can be fixed or discovered using a genetic operator called Module Mutation. Several versions of Module Mutation are evaluated in this paper. Both fixed modular networks and Module Mutation networks outperform monolithic networks and Multitask networks. Interestingly, the best networks dedicate modules to critical behaviors (such as escaping when surrounded after luring ghosts near a power pill) that do not follow the customary division of the game into chasing edible and escaping threat ghosts. The results demonstrate that MM-NEAT can discover interesting and effective behavior for agents in challenging games.

Ms. Pac-Man is a challenging video game in which multiple modes of behavior are required to succeed: Ms. Pac-Man must escape ghosts when they are threats, and catch them when they are edible, in addition to eating all pills in each level. Past approaches to learning behavior in Ms. Pac-Man have treated the game as a single task to be learned using monolithic policy representations. In contrast, this paper uses a framework called Modular Multiobjective NEAT to evolve modular neural networks. Each module defines a separate policy; evolution discovers these policies and when to use them. The number of modules can be fixed or learned using a new version of a genetic operator, called Module Mutation, which duplicates an existing module that can then evolve to take on a distinct behavioral identity. Both the fixed modular networks and Module Mutation networks outperform traditional monolithic networks. More interestingly, the best modular networks dedicate modules to critical behaviors that do not follow the customary division of the game into chasing edible and escaping threatening ghosts.

Generative Adversarial Networks (GANs) are an emerging form of indirect encoding. The GAN is trained to induce a latent space on training data, and a real-valued evolutionary algorithm can search that latent space. Such Latent Variable Evolution (LVE) has recently been applied to game levels. However, it is hard for objective scores to capture level features that are appealing to players. Therefore, this paper introduces a tool for interactive LVE of tile-based levels for games. The tool also allows for direct exploration of the latent dimensions, and allows users to play discovered levels. The tool works for a variety of GAN models trained for both Super Mario Bros. and The Legend of Zelda, and is easily generalizable to other games. A user study shows that both the evolution and latent space exploration features are appreciated, with a slight preference for direct exploration, but combining these features allows users to discover even better levels. User feedback also indicates how this system could eventually grow into a commercial design tool, with the addition of a few enhancements.

Abstract-Intelligent opponent behavior helps make video games interesting to human players. Evolutionary computation can discover such behavior, especially when the game consists of a single task. However, multitask domains, in which separate tasks within the domain each have their own dynamics and objectives, can be challenging for evolution. This paper proposes two methods for meeting this challenge by evolving neural networks: 1) Multitask Learning provides a network with distinct outputs per task, thus evolving a separate policy for each task, and 2) Mode Mutation provides a means to evolve new output modes, as well as a way to select which mode to use at each moment. Multitask Learning assumes agents know which task they are currently facing; if such information is available and accurate, this approach works very well, as demonstrated in the Front/Back Ramming game of this paper. In contrast, Mode Mutation discovers an appropriate task division on its own, which may in some cases be even more powerful than a human-specified task division, as shown in the Predator/Prey game of this paper. These results demonstrate the importance of both Multitask Learning and Mode Mutation for learning intelligent behavior in complex games.

No abstract

Abstract-The UTˆ2 bot, which had a humanness rating of 27.2727% in BotPrize 2010, is based on two core ideas: (1) multiobjective neuroevolution is used to learn skilled combat behavior, but filters on the available combat actions ensure that the behavior is still human-like despite being evolved for performance, and (2) a database of traces of human play is used to help the bot get unstuck when its navigation capabilities fail. Several changes have recently been made to UTˆ2: Extra input features have been provided to the bot to help it evolve better combat behavior, the role of human traces in the navigation of the bot has been expanded, and an extra control module has been added which encourages the bot to observe other players the way a human would, rather than simply battle them. These changes should make UTˆ2 act more human-like in this year's BotPrize competition.

No abstract