Reviewers all agree that the paper presents interesting theoretical relations between adversarial examples and explanations and is well-written. There are some suggestions on how to provide more intuitive explanations (no pun intended..) for proofs among other things. Please carefully incorporate feedback to your final version.