Skip to project
LIDOR.
← All projects
SOFTWARE & SYSTEMS

NEAT AUTONOMOUS GAMING AGENT

An Icy Tower clone built in PyGame, with a bot trained using the NEAT neuroevolution algorithm - play, train and test modes.

completed
Year
2020
Game
PyGame Icy Tower Clone
NEAT AGENT POPULATION TRAINING RUN

NEAT AGENT POPULATION TRAINING RUN

NEUROEVOLUTION OF AUGMENTING TOPOLOGIES IN ICY TOWER

NEATPYGAMENEUROEVOLUTIONGAME-AI
01 / The build

Behind the project.

A clone of the Icy Tower game built in PyGame, written as a platform for developing an AI bot: the bot is trained with NEAT (NeuroEvolution of Augmenting Topologies), and the game has play, train-bot and test-bot modes. A second iteration of the same game retrains the agent with Deep Q-learning (DQN) instead of NEAT (work in progress).

02 / In the workshop

Project media.

1 photos & media · Tap to expand

Technical details

Under the hood.

Game
PyGame Icy Tower Clone
AI
NEAT (python-neat)
Modes
Play / Train Bot / Test Bot
Notes & outcomes

What came out of it.

  • Neuroevolved game bot.
Keep exploring

Code & resources.