UCLA

Learning the structure of causal directed acyclic graphs is useful in many areas of machine learning and artificial intelligence, with applications in fields such as robotics, economics, and genomics. However, in the high-dimensional setting, it is challenging to obtain good empirical and theoretical results without strong and often restrictive assumptions. Additionally, it is questionable whether all of the variables purported to be included in the network are appropriate. It is of interest then to restrict consideration to a subset of the variables for relevant and reliable inferences. In fact, researchers in various disciplines can usually select a set of target nodes in the network for causal discovery. This dissertation develops a new constraint-based method for estimating the local structure around user-specified target nodes, employing rules from the Fast Causal Inference algorithm to coordinate structure learning between neighborhoods. Our method facilitates causal discovery without learning the entire DAG structure. We establish consistency results for our algorithm with respect to the local neighborhood structure of the target nodes in the true CPDAG. Empirical experimental results show that our algorithm is more accurate in learning the neighborhood structures with much less computational cost than standard methods that estimate the entire DAG. An R package implementing our algorithm may be accessed at http://github.com/stephenvsmith/CML.