Seth Cooper

Colan Biemer, Seth Cooper

An increasingly common area of study in procedural content generation is the creation of level segments: short pieces that can be used to form larger levels. Previous work has used basic concatenation to form these larger levels. However, even if the segments themselves are completable and well-formed, concatenation can fail to produce levels that are completable and can cause broken in-game structures (e.g. malformed pipes in Mario). We show this with three tile-based games: a side-scrolling platformer, a vertical platformer, and a top-down roguelike. Additionally, we present a Markov chain and a tree search algorithm that finds a link between two level segments, which uses filters to ensure completability and unbroken in-game structures in the linked segments. We further show that these links work well for multi-segment levels. We find that this method reliably finds links between segments and is customizable to meet a designer's needs.

Johor Jara Gonzalez, Seth Cooper, Matthew Guzdial

Automated game design (AGD), the study of automatically generating game rules, has a long history in technical games research. AGD approaches generally rely on approximations of human play, either objective functions or AI agents. Despite this, the majority of these approximators are static, meaning they do not reflect human player's ability to learn and improve in a game. In this paper, we investigate the application of Reinforcement Learning (RL) as an approximator for human play for rule generation. We recreate the classic AGD environment Mechanic Maker in Unity as a new, open-source rule generation framework. Our results demonstrate that RL produces distinct sets of rules from an A* agent baseline, which may be more usable by humans.

Anurag Sarkar, Seth Cooper

We present tile2tile, an approach for style transfer between levels of tile-based platformer games. Our method involves training models that translate levels from a lower-resolution sketch representation based on tile affordances to the original tile representation for a given game. This enables these models, which we refer to as filters, to translate level sketches into the style of a specific game. Moreover, by converting a level of one game into sketch form and then translating the resulting sketch into the tiles of another game, we obtain a method of style transfer between two games. We use Markov random fields and autoencoders for learning the game filters and apply them to demonstrate style transfer between levels of Super Mario Bros, Kid Icarus, Mega Man and Metroid.

Colan F. Biemer, Seth Cooper

Many games feature a progression of levels that doesn't adapt to the player. This can be problematic because some players may get stuck if the progression is too difficult, while others may find it boring if the progression is too slow to get to more challenging levels. This can be addressed by building levels based on the player's performance and preferences. In this work, we formulate the problem of generating levels for a player as a Markov Decision Process (MDP) and use adaptive dynamic programming (ADP) to solve the MDP before assembling a level. We tested with two case studies and found that using an ADP outperforms two baselines. Furthermore, we experimented with player proxies and switched them in the middle of play, and we show that a simple modification prior to running ADP results in quick adaptation. By using ADP, which searches the entire MDP, we produce a dynamic progression of levels that adapts to the player.

Venkata Sai Revanth Atmakuri, Seth Cooper, Matthew Guzdial

Game level blending via machine learning, the process of combining features of game levels to create unique and novel game levels using Procedural Content Generation via Machine Learning (PCGML) techniques, has gained increasing popularity in recent years. However, many existing techniques rely on human-annotated level representations, which limits game level blending to a limited number of annotated games. Even with annotated games, researchers often need to author an additional shared representation to make blending possible. In this paper, we present a novel approach to game level blending that employs Clustering-based Tile Embeddings (CTE), a learned level representation technique that can serve as a level representation for unannotated games and a unified level representation across games without the need for human annotation. CTE represents game level tiles as a continuous vector representation, unifying their visual, contextual, and behavioral information. We apply this approach to two classic Nintendo games, Lode Runner and The Legend of Zelda. We run an evaluation comparing the CTE representation to a common, human-annotated representation in the blending task and find that CTE has comparable or better performance without the need for human annotation.

Anurag Sarkar, Seth Cooper

We present latent combinational game design -- an approach for generating playable games that blend a given set of games in a desired combination using deep generative latent variable models. We use Gaussian Mixture Variational Autoencoders (GMVAEs) which model the VAE latent space via a mixture of Gaussian components. Through supervised training, each component encodes levels from one game and lets us define blended games as linear combinations of these components. This enables generating new games that blend the input games as well as controlling the relative proportions of each game in the blend. We also extend prior blending work using conditional VAEs and compare against the GMVAE and additionally introduce a hybrid conditional GMVAE (CGMVAE) architecture which lets us generate whole blended levels and layouts. Results show that these approaches can generate playable games that blend the input games in specified combinations. We use both platformers and dungeon-based games to demonstrate our results.

Mahsa Bazzaz, Seth Cooper

Determining the completability of levels generated by procedural generators such as machine learning models can be challenging, as it can involve the use of solver agents that often require a significant amount of time to analyze and solve levels. Active learning is not yet widely adopted in game evaluations, although it has been used successfully in natural language processing, image and speech recognition, and computer vision, where the availability of labeled data is limited or expensive. In this paper, we propose the use of active learning for learning level completability classification. Through an active learning approach, we train deep-learning models to classify the completability of generated levels for Super Mario Bros., Kid Icarus, and a Zelda-like game. We compare active learning for querying levels to label with completability against random queries. Our results show using an active learning approach to label levels results in better classifier performance with the same amount of labeled data.

Mahsa Bazzaz, Seth Cooper

Procedurally generated levels created by machine learning models can be unsolvable without further editing. Various methods have been developed to automatically repair these levels by enforcing hard constraints during the post-processing step. However, as levels increase in size, these constraint-based repairs become increasingly slow. This paper proposes using explainability methods to identify specific regions of a level that contribute to its unsolvability. By assigning higher weights to these regions, constraint-based solvers can prioritize these problematic areas, enabling more efficient repairs. Our results, tested across three games, demonstrate that this approach can help to repair procedurally generated levels faster.

Mahsa Bazzaz, Seth Cooper

Procedural content generation via machine learning (PCGML) in games involves using machine learning techniques to create game content such as maps and levels. 2D tile-based game levels have consistently served as a standard dataset for PCGML because they are a simplified version of game levels while maintaining the specific constraints typical of games, such as being solvable. In this work, we highlight the unique characteristics of game levels, including their structured discrete data nature, the local and global constraints inherent in the games, and the sensitivity of the game levels to small changes in input. We define the robustness of data as a measure of sensitivity to small changes in input that cause a change in output, and we use this measure to analyze and compare these levels to state-of-the-art machine learning datasets, showcasing the subtle differences in their nature. We also constructed a large dataset from four games inspired by popular classic tile-based games that showcase these characteristics and address the challenge of sparse data in PCGML by providing a significantly larger dataset than those currently available.

Mahsa Bazzaz, Seth Cooper

Expressive range analysis is a visualization-based technique used to evaluate the performance of generative models, particularly in game level generation. It typically employs two quantifiable metrics to position generated artifacts on a 2D plot, offering insight into how content is distributed within a defined metric space. In this work, we use the expressive range of a generator as the conceptual space of possible creations. Inspired by the quality diversity paradigm, we explore this space to generate levels. To do so, we use a constraint-based generator that systematically traverses and generates levels in this space. To train the constraint-based generator we use different tile patterns to learn from the initial example levels. We analyze how different patterns influence the exploration of the expressive range. Specifically, we compare the exploration process based on time, the number of successful and failed sample generations, and the overall interestingness of the generated levels. Unlike typical quality diversity approaches that rely on random generation and hope to get good coverage of the expressive range, this approach systematically traverses the grid ensuring more coverage. This helps create unique and interesting game levels while also improving our understanding of the generator's strengths and limitations.

Mahsa Bazzaz, Seth Cooper

Generative Adversarial Networks (GANs) are unsupervised models designed to learn and replicate a target distribution. The vanilla versions of these models can be extended to more controllable models. Conditional Generative Adversarial Networks (CGANs) extend vanilla GANs by conditioning both the generator and discriminator on some additional information (labels). Controllable models based on complementary learning, such as Rumi-GAN, have been introduced. Rumi-GANs leverage negative examples to enhance the generator's ability to learn positive examples. We evaluate the performance of two controllable GAN variants, CGAN and Rumi-GAN, in generating game levels targeting specific constraints of interest: playability and controllability. This evaluation is conducted under two scenarios: with and without the inclusion of negative examples. The goal is to determine whether incorporating negative examples helps the GAN models avoid generating undesirable outputs. Our findings highlight the strengths and weaknesses of each method in enforcing the generation of specific conditions when generating outputs based on given positive and negative examples.

Andrew Walter, Seth Cooper, Panagiotis Manolios

We describe the design and implementation of a reasoning engine that facilitates the gamification of loop-invariant discovery. Our reasoning engine enables students, computational agents and regular software engineers with no formal methods expertise to collaboratively prove interesting theorems about simple programs using browser-based, online games. Within an hour, players are able to specify and verify properties of programs that are beyond the capabilities of fully-automated tools. The hour limit includes the time for setting up the system, completing a short tutorial explaining game play and reasoning about simple imperative programs. Players are never required to understand formal proofs; they only provide insights by proposing invariants. The reasoning engine is responsible for managing and evaluating the proposed invariants, as well as generating actionable feedback.

Anurag Sarkar, Seth Cooper

Behavior trees (BTs) are a popular method for modeling NPC and enemy AI behavior and have been widely used in commercial games. In this work, rather than use BTs to model game playing agents, we use them for modeling game design agents, defining behaviors as content generation tasks rather than in-game actions. Similar to how traditional BTs enable modeling behaviors in a modular and dynamic manner, BTs for PCG enable simple subtrees for generating parts of levels to be combined modularly to form complex trees for generating whole levels as well as generators that can dynamically vary the generated content. We refer to this approach as Procedural Content Generation using Behavior Trees, or PCGBT, and demonstrate it by using BTs to model generators for Super Mario Bros., Mega Man and Metroid levels as well as dungeon layouts and discuss several ways in which this paradigm could be applied and extended in the future.

Anurag Sarkar, Seth Cooper

Variational autoencoders (VAEs) have been used in prior works for generating and blending levels from different games. To add controllability to these models, conditional VAEs (CVAEs) were recently shown capable of generating output that can be modified using labels specifying desired content, albeit working with segments of levels and platformers exclusively. We expand these works by using CVAEs for generating whole platformer and dungeon levels, and blending levels across these genres. We show that CVAEs can reliably control door placement in dungeons and progression direction in platformer levels. Thus, by using appropriate labels, our approach can generate whole dungeons and platformer levels of interconnected rooms and segments respectively as well as levels that blend dungeons and platformers. We demonstrate our approach using The Legend of Zelda, Metroid, Mega Man and Lode Runner.

Anurag Sarkar, Seth Cooper

Several works have demonstrated the use of variational autoencoders (VAEs) for generating levels in the style of existing games and blending levels across different games. Further, quality-diversity (QD) algorithms have also become popular for generating varied game content by using evolution to explore a search space while focusing on both variety and quality. To reap the benefits of both these approaches, we present a level generation and game blending approach that combines the use of VAEs and QD algorithms. Specifically, we train VAEs on game levels and run the MAP-Elites QD algorithm using the learned latent space of the VAE as the search space. The latent space captures the properties of the games whose levels we want to generate and blend, while MAP-Elites searches this latent space to find a diverse set of levels optimizing a given objective such as playability. We test our method using models for 5 different platformer games as well as a blended domain spanning 3 of these games. We refer to using MAP-Elites for blending as Blend-Elites. Our results show that MAP-Elites in conjunction with VAEs enables the generation of a diverse set of playable levels not just for each individual game but also for the blended domain while illuminating game-specific regions of the blended latent space.

Anurag Sarkar, Zhihan Yang, Seth Cooper

Prior research has shown variational autoencoders (VAEs) to be useful for generating and blending game levels by learning latent representations of existing level data. We build on such models by exploring the level design affordances and applications enabled by conditional VAEs (CVAEs). CVAEs augment VAEs by allowing them to be trained using labeled data, thus enabling outputs to be generated conditioned on some input. We studied how increased control in the level generation process and the ability to produce desired outputs via training on labeled game level data could build on prior PCGML methods. Through our results of training CVAEs on levels from Super Mario Bros., Kid Icarus and Mega Man, we show that such models can assist in level design by generating levels with desired level elements and patterns as well as producing blended levels with desired combinations of games.

Zhihan Yang, Anurag Sarkar, Seth Cooper

Variational autoencoders (VAEs) have been shown to be able to generate game levels but require manual exploration of the learned latent space to generate outputs with desired attributes. While conditional VAEs address this by allowing generation to be conditioned on labels, such labels have to be provided during training and thus require prior knowledge which may not always be available. In this paper, we apply Gaussian Mixture VAEs (GMVAEs), a variant of the VAE which imposes a mixture of Gaussians (GM) on the latent space, unlike regular VAEs which impose a unimodal Gaussian. This allows GMVAEs to cluster levels in an unsupervised manner using the components of the GM and then generate new levels using the learned components. We demonstrate our approach with levels from Super Mario Bros., Kid Icarus and Mega Man. Our results show that the learned components discover and cluster level structures and patterns and can be used to generate levels with desired characteristics.

Anurag Sarkar, Seth Cooper

In recent years, machine learning (ML) systems have been increasingly applied for performing creative tasks. Such creative ML approaches have seen wide use in the domains of visual art and music for applications such as image and music generation and style transfer. However, similar creative ML techniques have not been as widely adopted in the domain of game design despite the emergence of ML-based methods for generating game content. In this paper, we argue for leveraging and repurposing such creative techniques for designing content for games, referring to these as approaches for Game Design via Creative ML (GDCML). We highlight existing systems that enable GDCML and illustrate how creative ML can inform new systems via example applications and a proposed system.

Anurag Sarkar, Seth Cooper

Existing methods of level generation using latent variable models such as VAEs and GANs do so in segments and produce the final level by stitching these separately generated segments together. In this paper, we build on these methods by training VAEs to learn a sequential model of segment generation such that generated segments logically follow from prior segments. By further combining the VAE with a classifier that determines whether to place the generated segment to the top, bottom, left or right of the previous segment, we obtain a pipeline that enables the generation of arbitrarily long levels that progress in any of these four directions and are composed of segments that logically follow one another. In addition to generating more coherent levels of non-fixed length, this method also enables implicit blending of levels from separate games that do not have similar orientation. We demonstrate our approach using levels from Super Mario Bros., Kid Icarus and Mega Man, showing that our method produces levels that are more coherent than previous latent variable-based approaches and are capable of blending levels across games.

Anurag Sarkar, Zhihan Yang, Seth Cooper

Previous work explored blending levels from existing games to create levels for a new game that mixes properties of the original games. In this paper, we use Variational Autoencoders (VAEs) for improving upon such techniques. VAEs are artificial neural networks that learn and use latent representations of datasets to generate novel outputs. We train a VAE on level data from Super Mario Bros. and Kid Icarus, enabling it to capture the latent space spanning both games. We then use this space to generate level segments that combine properties of levels from both games. Moreover, by applying evolutionary search in the latent space, we evolve level segments satisfying specific constraints. We argue that these affordances make the VAE-based approach especially suitable for co-creative level design and compare its performance with similar generative models like the GAN and the VAE-GAN.

Zhengxing Chen, Truong-Huy D Nguyen, Yuyu Xu et al.

Multiplayer Online Battle Arena (MOBA) games have received increasing popularity recently. In a match of such games, players compete in two teams of five, each controlling an in-game avatars, known as heroes, selected from a roster of more than 100. The selection of heroes, also known as pick or draft, takes place before the match starts and alternates between the two teams until each player has selected one hero. Heroes are designed with different strengths and weaknesses to promote team cooperation in a game. Intuitively, heroes in a strong team should complement each other's strengths and suppressing those of opponents. Hero drafting is therefore a challenging problem due to the complex hero-to-hero relationships to consider. In this paper, we propose a novel hero recommendation system that suggests heroes to add to an existing team while maximizing the team's prospect for victory. To that end, we model the drafting between two teams as a combinatorial game and use Monte Carlo Tree Search (MCTS) for estimating the values of hero combinations. Our empirical evaluation shows that hero teams drafted by our recommendation algorithm have significantly higher win rate against teams constructed by other baseline and state-of-the-art strategies.

Zhengxing Chen, Chris Amato, Truong-Huy Nguyen et al.

Deck building is a crucial component in playing Collectible Card Games (CCGs). The goal of deck building is to choose a fixed-sized subset of cards from a large card pool, so that they work well together in-game against specific opponents. Existing methods either lack flexibility to adapt to different opponents or require large computational resources, still making them unsuitable for any real-time or large-scale application. We propose a new deck recommendation system, named Q-DeckRec, which learns a deck search policy during a training phase and uses it to solve deck building problem instances. Our experimental results demonstrate Q-DeckRec requires less computational resources to build winning-effective decks after a training phase compared to several baseline methods.