science1 distinct publisher
At ICML, authors ranking their own papers beat reviewer scores at spotting future citations
A year-long experiment covering 2,592 ICML 2023 submissions found top self-ranked papers drew about twice the citations of bottom-ranked ones. The signal survived rejection, which is the awkward part.
Publishers:nature.com
Reality
- Evidence74
- Adoption24
- Hype gap+12
- Incentives
- Insufficient
- Confidence56