Q-value

In reinforcement learning, a Q-value represents the expected future rewards that an agent can expect to receive by taking a specific action from a given state. It helps in determining the best action to take by comparing the Q-values of all possible actions from that state.