MirroS
Foundation Models & Learning
MirroS models both the agent and its environment: actions change the world, and the resulting feedback updates the agent. This dual loop addresses scarce interaction data and the gap between simulated and real physical-AI training.
