Show HN: Watch RL-trained squads play a 2D battle royale (full match replay)
Category: ai-ml
Tags: reinforcement-learning, multi-agent, game-ai, battle-royale, replay-viewer
Score: 6.3/10 (Innovation: 7, Technical: 8, Documentation: 6, Utility: 4)
survevBot trains squads of four agents to play a 2D battle royale using multi-agent reinforcement learning (MAPPO) and provides interactive replay viewer in the browser. The project is a proof of concept that demonstrates how RL can produce coordinated team behavior in a complex game environment, though training code and weights are withheld to prevent cheating.
Target audience: AI researchers, game developers, and RL enthusiasts
Repository: https://123wwwa.github.io/survevbot-demo/ · GPL-3.0 · 1 stars
View on Hacker News