IEEE COG 2021 - GVGAI Learning Competition


Organised by Hao Tong1, Tianye Shu1 and Jialin Liu1

Steering Committee: Jialin Liu1, Julian Togelius2, Diego Perez-Liebana3 and Simon M. Lucas3

1 Southern University of Science and Technology (SUSTech), China
2 New York University (NYU), USA
3 Queen Mary University of London (QMUL), UK

This GVGAI Single-Player Learning Competition is organised at the IEEE’s 2021 Conference on Games (CoG2021) .

More about the other GVGAI competition tracks (1-p planning, 2-p planning, level generation, rule generation) can be found here

Documents about GVGAI framework

  • A survey on IEEE Transactions on Games: General Video Game AI: a Multi-Track Framework for Evaluating Agents, Games and Content Generation Algorithms. -- by Diego Perez-Liebana, Jialin Liu, Ahmed Khalifa, Raluca D. Gaina, Julian Togelius, Simon M. Lucas.
  • A Morgan and Claypool Book: General Video Game Artificial Intelligence -- by Diego Perez-Liebana and Simon M. Lucas and Raluca D. Gaina and Julian Togelius and Ahmed Khalifa and Jialin Liu.

Competition Games

  • Games' Name: bravekeeper, greedymouse, trappedhero More details
  • Training Levels: Level 0 and Level 1
  • Testing Levels: Level 2, Level 3
  • Validation level: Level 4
  • The training levels have been embeded in the GVGAI Learning Framework. The testing and validation levels will be released after the competition.

  • With VGDL, you can generate additional levels by yourself to train your agent!

Competition Rules

  • Due to the long training time, the GVGAI server won’t be used for training your agent. Please train your agent using your own machine or server.
  • Only need to submitted the trained model.
  • The ranking board only provides the rank on testing levels.
  • Winners are based on the testing and validation levels!!

Other hints

1. Preparation

Download and set up the new GVGAI-Gym framework on your machine/server.

  • Docker:
    We have prepared the Dockerfile to help you set up the environment. Link

2. Training phase

Program your agent and train it

  • On as many games/levels as you want
  • Using as much time as you want for deciding an action per game tick
  • Using as much time as you want for training
  • On the training levels given by us

3. Test and Validate Phase

  • Unknown levels of the same games will be used for testing and validation.
  • You can find the script used for testing and validating on this webpage.

Submission

  • Compress your trained agent as a zip file.
  • Follow the submission rules to submit your agent. Submit
  • Deadline: 23:59 (GMT), July 31st, 2021

Game Description and Design

Game author: Tianye Shu

🎮 BraveKeeper

BraveKeeper is a Sokoban-like game, aiming at keeping the avatar and treasure chests away from monsters. The game is designed with a periodic reward that is given every 100 game ticks. Green Tiles are safe places where monsters can’t reach. You will lose score if the treasure chests are stolen by monsters. When all treasure chests are stolen or avatar collides with monsters, a game terminates with state PLAYER LOSES. A score will be rewarded if a treasure chest is pushed into safe places. If the avatar and any treasure chest survive for more than 1200 game ticks, then a game terminates with state PLAYER WINS.

To be added after competition
To be added after competition
To be added after competition
To be added after competition
To be added after competition
            Level 0
            Level 1
            Level 2
            Level 3
            Level 4

🎮 GreedyMouse

GreedyMouse is a resource-collection game. The avatar, thus the greedy mouse controlled by a player/agent, is expected to collect all the food in a level in which case a game terminates with state PLAYER WINS. However, some food is protected by the walls and there are traps in some food. Luckily, the greedy mouse has the ability to destroy the walls and check the trap. Don’t forget cats. The greedy mouse should avoid them when finding food. When a collision of the avatar with cats or traps, a game terminates with state PLAYER LOSES. And if maximum 1200 game ticks is passed and greedy mouse does not eat all food, a game terminates with state PLAYER LOSES.

To be added after competition
To be added after competition
To be added after competition
To be added after competition
To be added after competition
            Level 0
            Level 1
            Level 2
            Level 3
            Level 4

🎮 Trappedhero

TrappedHero is a maze game with sparse reward. There are only three sprites in each level: the avatar, an immobile key and an immobile door. Scores are rewarded in two situations: the avatar collects the key or the avatar touches the door with a key. The latter will end the game with state PLAYER WINS, otherwise, the game terminates with state PLAYER LOSES after 1200 game ticks.

To be added after competition
To be added after competition
to be added
To be added after competition
To be added after competition
            Level 0
            Level 1
            Level 2
            Level 3
            Level 4


Affiliations