In machine learning, an agent refers to a system that is capable of making decisions or taking actions in an environment to achieve a specific goal. The very term "agent" emphasizes that the system operates independently and is capable of making decisions based on its own observations and actions, without human involved.

For example, a self-driving car can be considered as an agent, which goal is to safely transport passengers from point A to point B. The agent uses sensors and cameras to perceive its environment, and algorithms to make decisions on how to control the car's movement, such as accelerating, braking, and steering. The agent continually gets better at making these decisions because it learns from its experiences and the data it collects.