幂级智能
基础模型与学习范式
幂级智能以世界模型和密集强化学习研究具身智能，使自动驾驶与具身系统能够生成高价值训练数据，并更有效地进行测试和训练。
DenseAI combines world models with dense reinforcement learning for embodied AI. The approach aims to generate useful training data and scale testing for autonomous driving and embodied systems.
