Multi-Domain Level Generation and Blending with Sketches via Example-Driven BSP and Variational Autoencoders

Snodgrass, Sam; Sarkar, Anurag

doi:10.1145/3402942.3402948

Cited by 24 publications

(32 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Finally, in generating levels blended across different games, our work is situated amidst a recent line of PCGML research focusing on more creative applications of ML for game design [19,33]. Such techniques touch upon concepts of combinational creativity [2] and have included domain transfer [39,42], automated game generation [18] and game blending [13] which refers to generating new games by combining properties such as levels and/or mechanics of existing games. While several recent works have demonstrated learning to blend games using the VAE latent space [34,35], using MAP-Elites could help produce a wide variety of blended game levels as well as identify if certain regions of the blended latent space correspond to specific combinations of games being blended.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Generating and Blending Game Levels via Quality-Diversity in the Latent Space of a Variational Autoencoder

Sarkar

Cooper

2021

The 16th International Conference on the Foundations of Digital Games (FDG) 2021

Self Cite

View full text Add to dashboard Cite

Several recent works have demonstrated the use of variational autoencoders (VAEs) for both generating levels in the style of existing games as well as blending levels across different games. Additionally, 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. In order 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 then 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. Our results show that using 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.

show abstract

Section: Related Workmentioning

confidence: 99%

“…Thus, the maximum value for this is 32 since there are 16 rows and 16 columns. As a metric, it is similar to the plagiarism metrics as defined in [42,44]. Thus, the ranges for Symmetry and Similarity were [0, 256] and [0, 32] respectively, giving us an archive consisting of 257 × 33 = 8,481 cells.…”

Section: Map-elitesmentioning

confidence: 99%

Generating and Blending Game Levels via Quality-Diversity in the Latent Space of a Variational Autoencoder

Sarkar

Cooper

2021

The 16th International Conference on the Foundations of Digital Games (FDG) 2021

Self Cite

View full text Add to dashboard Cite

show abstract

“…The method presented in this paper allows generating whole level blends in a controllable manner in addition to blending games from different genres. In doing so, this work constitutes another PCGML approach under combinational creativity [24], similar to recent works looking at domain transfer [25] and automated game generation [26], all of which blend and/or combine existing domains to generate content for new ones. Additionally, in being a creative ML approach for level generation and blending, this work can also be categorized under the recently defined Game Design via Creative Machine Learning (GDCML) [27] set of approaches.…”

Section: Introductionmentioning

confidence: 99%

Dungeon and Platformer Level Blending and Generation using Conditional VAEs

Sarkar¹,

Cooper²

2021

Preprint

Self Cite

View full text Add to dashboard Cite

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.

show abstract

“…In this vein, a recent trend of more creative PCGML has emerged [44], focusing on applications such as domain transfer [45], [46], level and game blending [47]- [49] and generation of entire new games using ML models [50]. These new works combined with the emergence of new ML-powered game design tools [51] signal that creative ML approaches prevalent primarily in visual art and music thus far, can be repurposed for use with ML models for game design.…”

Section: Introductionmentioning

confidence: 99%

Towards Game Design via Creative Machine Learning (GDCML)

Sarkar

Cooper

2020

2020 IEEE Conference on Games (CoG)

Self Cite

View full text Add to dashboard Cite

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.

show abstract

Multi-Domain Level Generation and Blending with Sketches via Example-Driven BSP and Variational Autoencoders

Cited by 24 publications

References 12 publications

Generating and Blending Game Levels via Quality-Diversity in the Latent Space of a Variational Autoencoder

Generating and Blending Game Levels via Quality-Diversity in the Latent Space of a Variational Autoencoder

Dungeon and Platformer Level Blending and Generation using Conditional VAEs

Towards Game Design via Creative Machine Learning (GDCML)

Contact Info

Product

Resources

About