ICLR Breach | The Pitfall of Peer Review
Another vivid example of “in the worst case, what (tf) would happen if xxx information is leaked?” ICLR is one of the most important computer science conferences. They do double-blind peer review on OpenReview. OpenReview makes all paper submissions and reviews public (e.g. anyone can view and make public comment on it, like Twitter). Reviewers goes under anonymous encrypted alias. Around Nov 11, a bug was found in OpenReview that you can query the anonymous encrypted alias of reviewers and get the true reviewer identity. Consider the openness of OpenReview, technically, everything was under the daylight from Nov 11 to Nov 27 (when at Thankgiving morning, the officials learnt and fixed the bug). ...