Eligibility Traces

Fact

A reinforcement learning memory for giving recent moves credit after delayed rewards.

In Plain Words

Eligibility traces are fading sticky notes on an AI's recent moves. A late reward reads the newest notes first. Old notes get tiny cookie crumbs.

You see them in TD Learning and Actor-Critic. They help late rewards travel back faster.

Related Concepts

• Temporal-Difference Learning
• Reinforcement Learning
• Actor-Critic

TD Learning
Eligibility traces let TD Learning give reward credit to recent steps.

RL
Eligibility traces help RL learn from late rewards.

Actor-Critic
Actor-Critic can use eligibility traces to update its policy and value.