Overview
A learning agent selects actions and improves a policy using reward feedback. Robotics applications must account for physical safety and differences between training environments and deployment. G1’s manufacturer describes imitation and reinforcement learning.[1][2]
References
- Reinforcement Learning: An IntroductionMIT Press · Other · Published 2018-11-13 · Accessed 2026-10-03↑ Return to article
- Unitree G1 specificationsUnitree Robotics · Manufacturer · Publication date not stated · Accessed 2026-10-03↑ Return to article