Goto

Collaborating Authors

 posterior inclusion probability


94130ea17023c4837f0dcdda95034b65-AuthorFeedback.pdf

Neural Information Processing Systems

In the revision, we shall add the following table on average computational times of all the18 methodsbasedon10replications. Without the Bayesian32 machinery in the paper, we cannot extract the posterior inclusion probabilities for structure recovery using (2.5)33 and provide consequent strong guarantees for graph selection in Theorem 3. iii) For scalability, we compute the34 MAP estimator instead of sampling from the full posterior.