| Term | One-liner |
|---|---|
| Manipulation / Locomotion | Acting on the world vs moving the body |
| Policy | Mapping from observations to actions |
| IL / BC | Imitation learning / behavior cloning |
| RL | Reinforcement learning |
| VLA / VA | Vision-Language-Action / Vision-Action only |
| World Model / WAM | World model / World-Action Model |
| Teleoperation | Human remote control for demos |
| Sim2Real / Real2Sim | Sim→real transfer / real→sim reconstruction |
| Affordance | How an object can be acted upon |
| DoF / End-effector | Degrees of freedom / tool at the arm tip |
| FK / IK | Forward / inverse kinematics |
| URDF | Robot description format |
| MPC / WBC | Model predictive control / whole-body control |
| SLAM | Simultaneous localization and mapping |
Adapted from Embodied-AI-Guide’s terminology table.