Your task is to program an intelligent agent navigating a 2D grid world. Your agent controls a team of units collecting powerups and placing explosives, with the ultimate goal of taking your opponent down.
Bomberland is a challenging problem for out-of-the-box machine learning algorithms. Be prepared to manage real-time decision making, planning, game theory, and both adversarial and cooperative play.
Bomberland will feature an ongoing, always-on arena with an active leaderboard. Participants can get direct feedback on their strategies in 1v1 matches against other players.
From time to time, we’ll hold tournaments featuring live streams and prizes. Check out our previous AI Sports Challenge streams for a sneak peek of what’s ahead.
We’re creating Bomberland as a place for the community to explore and experiment with the latest cutting edge technologies from tree search algorithms to deep reinforcement learning.
You’ll want to check it out if:
- You’re looking for a challenging hands-on ML project
- You’re looking for a place to try out new libraries, frameworks, or research papers
- You’ve been fascinated by the work of companies like DeepMind and OpenAI
We envision Bomberland to evolve over time with new metas and challenges.
Bomberland is part of our larger goal at Coder One to make cutting-edge ML accessible. We’re focused on building out the right tools and infrastructure to support the community in progressively pushing the boundaries of what’s possible.
The Bomberland competition is now live 🎉
We have starter kits in Python and TypeScript to help you get started (and encourage any community contributions to the starter kit repo).