Are ICLR 2027 reviews unusually shallow this year, or is it just me? [D]
retained text
I’ve been going through ICLR 2027 reviews, and most of them look extremely shallow. They’re so f short, like they're written for a random workshop. Almost every review I've seen is two generic points under each section (strengths, weaknesses, questions) and nothing more. Then a score of 2, even when the issues they raise would justify a straight reject. There’s almost no discussion of the method's motivation, impact, or soundness, and they take it for granted. Am I exaggerating, or are you seeing the same thing? Because I'm really questioning the point of PhD or research generally speaking, more than ever
document fetched 2026-11-08 00:24 · canonical https://reddit.com/r/MachineLearning/comments/1xq1f1m/are_iclr_2027_reviews_unusually_shallow_this_year