You are now in control of a car stuck in a valley between two hills. Your goal is to reach the top of the hill on the right.
In reinforcement learning, agents must explore their environment to learn how to take actions. Try pressing different keys to explore how the car moves.
Objective: Figure out how to move the car forward and backward.
Now that you know how to move, your task is to reach the goal. This represents the idea of a goal state in reinforcement learning, where an agent must learn how to achieve an objective.
Objective: Your goal is to reach the flag at the top of the hill.
Every action you take now comes with a cost. In reinforcement learning, agents receive rewards or penalties for actions and learn to optimise their behavior to achieve goals efficiently.
Reward System: A reward of -1 for each timestep
Objective: Reach the flag while minimising penalties.
Amazing! You completed the challenge
You’ve just experienced firsthand what it’s like to be an agent learning in a complex environment. This process conceptually mirrors how reinforcement learning agents actually work.
Curious about how reinforcement learning works behind the scenes? Click the "Learn More" button to find out.