The Nature of Temporal Difference Errors in Multi-step Distributional Reinforcement Learning.
Yunhao TangRémi MunosMark RowlandBernardo Ávila PiresWill DabneyMarc G. BellemarePublished in: NeurIPS (2022)
Keyphrases
- temporal difference
- multi step
- td learning
- reinforcement learning
- function approximation
- model free
- reinforcement learning algorithms
- evaluation function
- temporal difference learning
- monte carlo
- function approximators
- action selection
- policy evaluation
- step size
- knn
- state space
- temporal difference methods
- k nearest neighbor
- policy iteration
- supervised learning
- actor critic
- neural network
- markov decision processes
- reinforcement learning methods
- training samples
- semi supervised
- machine learning
- data sets