Policy-Based Bayesian Active Causal Discovery with Deep Reinforcement Learning.
Heyang GaoZexu SunHao YangXu ChenPublished in: KDD (2024)
Keyphrases
- causal discovery
- reinforcement learning
- causal models
- optimal policy
- bayesian networks
- belief nets
- policy search
- causal relationships
- markov decision process
- action selection
- causal structure
- state space
- control policy
- action space
- function approximation
- discovery process
- reward function
- markov decision processes
- directed acyclic graph
- policy gradient
- markov blanket
- conditional independence
- partially observable
- decision problems
- causal relations
- dynamic programming
- partially observable markov decision processes
- function approximators
- reinforcement learning algorithms
- sufficient conditions
- bayesian inference
- maximum likelihood
- learning algorithm
- machine learning
- influence diagrams
- probability distribution
- temporal difference
- decision making
- data mining
- data analysis