Museum of Failures
Honest archive · 05
Museum of Failures
Every rejection is a draft of a better paper. Here are mine.
7
Rejections
3
Acceptances
30%
Hit rate
- 2023-09CVPR · RejectReviewers liked the idea, hated the experiments. Fair.
- 2023-11AAAI · RejectBorderline. We ran it again with more seeds — it held.
- 2024-03ICML · RejectStrong reject from R3. We rewrote the introduction completely.
- 2024-05ECCV · RejectOff topic for the venue, in retrospect.
- 2024-08NeurIPS · RejectReviewers disagreed strongly. We picked the harshest reviewer's framing — it improved the paper.
- 2024-11AAAI · AcceptSame paper, different audience.
- 2025-03CVPR · RejectTwo strong accepts, one strong reject. The lottery was unkind.
- 2025-05NeurIPS · AcceptThe ablations finally told the story end-to-end.
- 2025-09ICLR · RejectHonestly, we submitted too early.
- 2026-01ICML · AcceptTwo years and four versions later.