CVPR Paper Controversy; ML Community Reviews Peer Review

“ There is no evidence so far of major ML conferences adopting any such software programs.” Let me provide some evidence. All of the major machine learning conferences employ the Toronto Paper Matching system, which compares each submission to papers published by the reviewers. NIPS applies a fairly sophisticated model to normalize reviewer scores. I’m not sure about plagiarism detection, but all submissions to arXiv (which covers many submissions to these conferences) are checked for plagiarism.