IEEE CIS Games Technical Committee

                   --- Task Force on Procedural Content Generation


Chair: Dr. Jialin Liu, Southern University of Science and Technology (SUSTech), China

Vice Chair: Prof. Georgios Yannakais, University of Malta, Malta

Vice Chair: Prof. Simon M. Lucas, Queen Mary University of London (QMUL), UK


  • Chengpeng Hu, Southern University of Science and Technology (SUSTech), China
  • Ziqi Wang, Southern University of Science and Technology (SUSTech), China
  • Yuchen Li, Southern University of Science and Technology (SUSTech), China
  • Marco Scirea, University of Southern Denmark (SDU), Denmark
  • Mike Preuss, Leiden University (LEI), Holland
  • Cameron Browne, Maastricht University (UM), Holland
  • Mark Nelson, American University (AU), America

Benchmark platforms

  • General Video Game AI (GVGAI) framework is an OpenAI Gym environment for games written in the Video Game Description Language (VGDL), including the GVGAI framework:
  • The Mario AI framework is a framework for using AI methods with a version of Super Mario Bros which was created by Ahmed Khalifa, based on the original Mario AI Framework by Sergey Karakovskiy, Noor Shaker, and Julian Togelius, which in turn was based on Infinite Mario Bros by Markus Persson [ code ]:
    • Sergey Karakovskiy and Julian Togelius, ''The Mario AI Benchmark and Competitions ,'' IEEE Transactions on Computational Intelligence and AI in Games (TCIAG), volume 4 issue 1, 55-67. [ pdf ]

Demo Code

  • CNet-assisted Evolutionary Level Repairer: This project demonstrates how defective game levels generated by machine learning models can be repaired through an evolutionary algorithm and neural network, without rules and domain knowledge. [ Code ]
  • Latent Space Illumination: This project implements a method for Latent Space Illumination (LSI) which explores the latent space of generative adversarial network via modern quality diversity algorithms. The outcome of LSI is a wide range of quality levels with distinct attributes (e.g., number of enemies & number of sky tiles). [ Code ]
  • PCGRL: An OpenAI GYM environment for Procedural Content Generation via Reinforcement Learning (PCGRL). This project demonstrates multiple ways to model offline game level generation tasks into reinforcement learning environments. Some trained reinforcement learning agents are also included. [ Code ]
  • MarioGPT: MarioGPT is a fine-tuned GPT2 model (specifically, distilgpt2), trained on a subset of Super Mario Bros and Super Mario Bros: The Lost Levels levels, provided by The Video Game Level Corpus. MarioGPT is able to generate levels, guided by a simple text prompt. [ Code ]
  • MarioPuzzle: This project is the official implementation of the online game level generation framework called "Experience-Driven PCG via Reinforcement Learning (EDRL)." Online level generation tasks are modeled as a Markov decision process and implemented as OpenAI GYM environments in this project. This project also implements a deep reinforcement learning algorithm to train agents for the purpose of generating fun and playable levels in real-time. [ Code ]
  • MarioGAN: This project allows for the unsupervised learning of a Generative Adversarial Network (GAN) that understands the structure of Super Mario Bros. levels. The model is trained on actual Mario levels from the Video Game Level Corpus. The trained model is capable of generating new level segments with the input of a latent vector, and these segments can be stitched together to make complete levels. [ Code ]

Books, Journal Special Issues and Review Papers

Organised Symposium, Workshops, and Special Sessions


Dr. Jialin Liu
Department of Computer Science and Engineering
Southern University of Science and Technology

College of Engineering, South Tower, Room 514
1088 Xueyuan Ave, Nanshan District
Shenzhen, Guangdong Province, China, 518055