Optimal experimental design via Bayesian optimization: active causal structure learning for Gaussian process networks.
Julius von KügelgenPaul K. RubensteinBernhard SchölkopfAdrian WellerPublished in: CoRR (2019)
Keyphrases
- structure learning
- experimental design
- gaussian process
- bayesian networks
- gaussian processes
- sample size
- hyperparameters
- graphical models
- approximate inference
- marginal likelihood
- model selection
- regression model
- network structure
- bayesian framework
- semi supervised
- active learning
- conditional independence
- probabilistic model
- latent variables
- posterior probability
- closed form
- cross validation
- belief propagation
- markov networks
- posterior distribution
- parameter estimation
- social networks
- optimal solution
- bayesian inference
- conditional probabilities
- higher order
- worst case
- random sampling
- class imbalance
- data mining
- random variables
- maximum likelihood
- em algorithm
- machine learning