Return to Dr. Jacob Schrum's Homepage

Generating Super Mario Bros Levels With Text-Conditional Diffusion Models

This page presents research in Procedural Content Generation via Machine Learning done in collaboration with undergraduate students Olivia Kilday, Bess Hagan, Emilio Salas, and Reid Williams as part of Southwestern University's Summer research program.

Source code for training your own Mario Diffusion models is available on GitHub. The GitHub repository also has simple instructions for downloading our models from Hugging Face and generating your own Mario levels without even training a model!

Out publication has been accepted to The 21st AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment (AIIDE 2025), but an arXiv pre-print explaining our approach and results is available here. The online appendix for the paper is included in the arXiv version, but is available separately at this link.

Video Demonstrating Our Interactive GUI For Level Creation


Model-Generated Scenes


Each collection of images shows level-scenes generated from trained models. The input-prompt for each image is shown. The same prompts are used for each model. The first 5 prompts are taken from real data, and the next 5 are randomly generated prompts not in the original data set. The first number in each image filename is simply to specify the order: 0 through 9. The second number in each filename is the caption adherence score: the degree to which the generated image actually matches the caption that is shown in the filename. Caption adherence scores range from -1.0 to 1.0, with 1.0 being a perfect score.

Regular Captions

The caption shown is the input the the diffusion model

MLM-regular0

MiniLM-single-regular0

MiniLM-multiple-regular0

GTE-single-regular0

GTE-multiple-regular0


Absence Captions

Though not shown in the file names, the actual input prompt for each image also includes an explicit mention of each feature that is completely absent from the scene. For example, prompts would contain phrases such as "no enemies. no pipes. no rectangular block clusters."

MLM-absence0

MiniLM-single-absence0

MiniLM-multiple-absence0

GTE-single-absence0

GTE-multiple-absence0


Negative Captions

Though not shown in the file names, each of these input prompts were combined with a separate corresponding negative prompt. So, if the input prompt did not make any mention of coins or pipes, then the corresponding negative prompt would include "coins. pipes." within its collection of phrases.

MLM-negative0

MiniLM-single-negative0

MiniLM-multiple-negative0

GTE-single-negative0

GTE-multiple-negative0

Related Publications


Peer-Reviewed Journal Articles


Peer-Reviewed Conference Publications


Extended Abstracts


Technical Reports


Dagstuhl Reports


Undergraduate Poster Presentations Supervised


Related Movies and Images


Miscellaneous Content

  • Summer 2025: Procedural Content Generation Using Generative AI: SURF Research Presentation made by my SURF Summer research students to present to other SURF students
  • Summer 2021: Quality Diversity and Creative Divergent Search: SCOPE Research Presentation made by my SCOPE Summer research students to present to other SCOPE students
  • Fall 2020: Procedural Content Generation for Games with Generative Adversarial Networks: Presentation for the Games AI Research Group at Queen Mary University of London (video)
  • Fall 2020: Interactively Evolving Video Game Levels with Generative Adversarial Networks: 403 Lecture for Math and CS Department
  • Fall 2020: Evolving Mega Man Levels with Generative Adversarial Networks: Virtual SCOPE Open House Website
  • Fall 2020: Evolving Lode Runner Levels with Generative Adversarial Networks: Virtual SCOPE Open House Website
  • Summer 2020: Generating Video Game Levels Using AI: SCOPE Research Presentation made by my SCOPE Summer research students to present to other SCOPE students
  • Summer 2019: Playing and Creating Games With Deep Neural Networks: "Mad Science Monday" presentation made by my SCOPE Summer research students to present to other SCOPE students
  • Fall 2018: Evolutionary Computation Applied to Digital Entertainment and the Arts, poster presented at the President's Appreciation Celebration for Southwestern University donors.
  • Fall 2018: Evolving Mario Levels in the Latent Space of a Deep Convolutional Generative Adversarial Network: presented to Southwestern University students as a 107 Lecture.
  • Summer 2018: The machines have taught themselves to make Mario levels, article in Fast Company about my recent GECCO 2018 paper on generating levels for Super Mario Bros.
  • Spring 2018: Bored with your video game? Artificial intelligence could create new levels on the fly, article in Science about my recent GECCO 2018 paper on generating levels for Super Mario Bros.
  • Spring 2018: Doom and Super Mario could be a lot tougher now AI is building levels, article in The Register about my recent GECCO 2018 paper on generating levels for Super Mario Bros.

  • Last Updated: 8/14/2025