Conditional Level Generation and Game Blending

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.

[1]  Yoshua Bengio,et al.  Generative Adversarial Nets , 2014, NIPS.

[2]  Matthew Guzdial,et al.  Learning to Blend Computer Game Levels , 2016, ICCC.

[3]  Honglak Lee,et al.  Learning Structured Output Representation using Deep Conditional Generative Models , 2015, NIPS.

[4]  Jichen Zhu,et al.  Graph grammar-based controllable generation of puzzles for a learning game about parallel programming , 2017, FDG.

[5]  Simon M. Lucas,et al.  Evolving mario levels in the latent space of a deep convolutional generative adversarial network , 2018, GECCO.

[6]  Max Welling,et al.  Auto-Encoding Variational Bayes , 2013, ICLR.

[7]  Julian Togelius,et al.  Multiobjective exploration of the StarCraft map space , 2010, Proceedings of the 2010 IEEE Conference on Computational Intelligence and Games.

[8]  Julian Togelius,et al.  DeepMasterPrints: Generating MasterPrints for Dictionary Attacks via Latent Variable Evolution* , 2017, 2018 IEEE 9th International Conference on Biometrics Theory, Applications and Systems (BTAS).

[9]  Santiago Ontañón,et al.  Controllable Procedural Content Generation via Constrained Multi-Dimensional Markov Chain Sampling , 2016, IJCAI.

[10]  Anurag Sarkar,et al.  Towards Game Design via Creative Machine Learning (GDCML) , 2020, 2020 IEEE Conference on Games (CoG).

[11]  Michael Mateas,et al.  Rhythm-based level generation for 2D platformers , 2009, FDG.

[12]  Honglak Lee,et al.  Attribute2Image: Conditional Image Generation from Visual Attributes , 2015, ECCV.

[13]  Jacob Schrum,et al.  Interactive evolution and exploration within latent level-design space of generative adversarial networks , 2020, GECCO.

[14]  Julian Togelius,et al.  Sentient Sketchbook: Computer-aided game level authoring , 2013, FDG.

[15]  Jacob Schrum,et al.  Generative Adversarial Network Rooms in Generative Graph Grammar Dungeons for The Legend of Zelda , 2020, 2020 IEEE Congress on Evolutionary Computation (CEC).

[16]  Julian Togelius,et al.  Interactive Constrained MAP-Elites Analysis and Evaluation of the Expressiveness of the Feature Dimensions , 2020, ArXiv.

[17]  Maria L. Rizzo,et al.  Energy statistics: A class of statistics based on distances , 2013 .

[18]  Anurag Sarkar,et al.  Controllable Level Blending between Games using Variational Autoencoders , 2020, ArXiv.

[19]  Julian Togelius,et al.  Procedural Content Generation via Machine Learning (PCGML) , 2017, IEEE Transactions on Games.

[20]  Anurag Sarkar,et al.  Blending Levels from Different Games using LSTMs , 2018, AIIDE Workshops.

[21]  Adam Summerville,et al.  Expanding Expressive Range: Evaluation Methodologies for Procedural Content Generation , 2018, AIIDE.

[22]  Gilles Fauconnier,et al.  Conceptual Integration Networks , 1998, Cogn. Sci..

[23]  Anurag Sarkar,et al.  Exploring Level Blending across Platformers via Paths and Affordances , 2020, AIIDE.

[24]  Santiago Ontañón,et al.  The VGLC: The Video Game Level Corpus , 2016, ArXiv.

[25]  Joris Dormans,et al.  Adventures in level design: generating missions and spaces for action adventure games , 2010, PCGames@FDG.

[26]  Anurag Sarkar,et al.  Multi-Domain Level Generation and Blending with Sketches via Example-Driven BSP and Variational Autoencoders , 2020, FDG.

[27]  Michael Mateas,et al.  Super Mario as a String: Platformer Level Generation Via LSTMs , 2016, DiGRA/FDG.

[28]  Joseph Corneli,et al.  Towards Generating Novel Games Using Conceptual Blending , 2021, Proceedings of the AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment.

[29]  Santiago Ontañón,et al.  Learning to Generate Video Game Maps Using Markov Models , 2017, IEEE Transactions on Computational Intelligence and AI in Games.

[30]  Anurag Sarkar,et al.  Sequential Segment-based Level Generation and Blending using Variational Autoencoders , 2020, FDG.

[31]  Julian Togelius,et al.  Bootstrapping Conditional GANs for Video Game Level Generation , 2019, 2020 IEEE Conference on Games (CoG).

[32]  Matthew Guzdial,et al.  Game Level Generation from Gameplay Videos , 2021, AIIDE.

[33]  Julian Togelius,et al.  Patterns and procedural content generation: revisiting Mario in world 1 level 1 , 2012, DPG '12.

[34]  Simon Osindero,et al.  Conditional Generative Adversarial Nets , 2014, ArXiv.

[35]  Luca Antiga,et al.  Automatic differentiation in PyTorch , 2017 .