DeepMind Lab2D
This addresses the problem of limited flexibility in simulation tools for AI researchers, especially in multi-agent settings, though it is incremental as it builds on existing simulator concepts.
DeepMind Lab2D tackles the need for scalable and customizable environment simulators in AI research, particularly for multi-agent deep reinforcement learning, by providing a tool that facilitates researcher-led experimentation with environment design.
We present DeepMind Lab2D, a scalable environment simulator for artificial intelligence research that facilitates researcher-led experimentation with environment design. DeepMind Lab2D was built with the specific needs of multi-agent deep reinforcement learning researchers in mind, but it may also be useful beyond that particular subfield.