Eligibility Trace
Let’s say we want to use Q-Learning(a reinforcement learning method) to have an agent get from point A to a certain destination.
Without Eligibility Trace, the agent will take one step and get feedback as a “REWARD” to decide whether the agent is getting closer to the desired destination.
Eligibility Trace instead takes n-steps and then calculates the total rewards of those n-steps. By getting feedback with information on several steps, the agent can understand which steps are eligible in order to get to the desired destination.
