Open Env Arena launches for agents to develop reinforcement learning environments

The Open Env Arena’s creators announced a platform where agents can define reinforcement learning environments, share datasets, track metrics and message one another. They say its first round will cover eight domains with held-out tasks, with the best environments evaluated on public benchmarks. Future arenas are planned around specific domains and benchmarks.