bd33f02c4e28615b5af2d24703e066d5-AuthorFeedback.pdf

Response to Reviewer #3: We think that the overall relevance of the article comes from the combination of the25 positive result ("if the degree assumptions on the hidden nodes are verified, then we can successfully recover the26 model") with thenegativeresult ("ifthedegree assumptions arenotverified, then there isapolytree model with the27 same observable nodes and fewer hidden nodes"). When the degree assumptions are not met, our method selects a28 polytree with aminimal number ofhidden nodes.