Get the latest tech news
Queueing to publish in AI and CS
Welcome to my academic site.
Here, I analyze some interesting properties that model the reviewing and acceptance system of machine learning conferences, but applies to CS more generally. The simulations yield that a decrease in the rate of acceptance from 35% to 20% increases the number of abandoned bad papers comparably to changing their wait in the pool for two or three rounds. One can observe other natural phenomena, such as a great reduction in the absolute number of average papers that were rejected merely due to bad luck wrt the others, when $\p$ increases to $0.35$.
Or read this on Hacker News