TD or not TD: Analyzing the Role of Temporal Differencing in Deep Reinforcement Learning.
Artemij AmiranashviliAlexey DosovitskiyVladlen KoltunThomas BroxPublished in: CoRR (2018)
Keyphrases
- reinforcement learning
- temporal difference
- reinforcement learning algorithms
- temporal difference learning
- eligibility traces
- td learning
- function approximation
- evaluation function
- policy evaluation
- learning algorithm
- state space
- model free
- markov decision processes
- monte carlo
- fixed point
- supervised learning
- action selection
- reinforcement learning methods
- machine learning
- control problems
- real time
- multi agent
- step size
- neural network
- policy iteration
- decision making
- case study
- game playing
- least squares