TD Learning
A class of model-free methods in reinforcement learning that combine ideas from dynamic programming and Monte Carlo methods, updating value estimates through bootstrapping.
Timeline (last 90 days)
Eligibility traces are the bridge connecting Monte Carlo and TD learning: a "trace" that decays over time records how recently each state was visited, and the λ parameter controls the degree of interpolation, making it the mechanistic basis of algorithms such as TD(λ) and SARSA(λ).
The paper migrates and adapts the reverse-time hitting-cost compensation technique from the analysis of TD learning to the state-action propagation setting.
All Facts (2)
Eligibility traces are the bridge connecting Monte Carlo and TD learning: a "trace" that decays over time records how recently each state was visited, and the λ parameter controls the degree of interpolation, making it the mechanistic basis of algorithms such as TD(λ) and SARSA(λ).
50%UnverifiedThe paper migrates and adapts the reverse-time hitting-cost compensation technique from the analysis of TD learning to the state-action propagation setting.
50%