Causal variables from reinforcement learning using generalized Bellman equations.
Tue HerlauPublished in: CoRR (2020)
Keyphrases
- reinforcement learning
- causal relationships
- causal models
- causal relations
- causal structure
- function approximation
- bayesian networks
- temporal difference learning
- causal bayesian networks
- structural model
- directed acyclic graph
- linear program
- markov decision processes
- markov blanket
- actor critic
- causal discovery
- mathematical model
- causal graph
- state space
- dynamic programming
- temporal difference
- multi agent
- numerical solution
- control problems
- learning algorithm
- causal reasoning
- hidden variables
- variable selection
- random variables
- machine learning
- linear equations
- reinforcement learning algorithms
- reward function
- piecewise linear
- state action
- experimental data
- optimal policy
- learning process
- causal knowledge
- causal ordering