Figure 1b) shows >= 90% of #NeurIPS2022 submissions thought they had a >= 25% chance of getting accepted. But only ~25% were accepted. Does this mean:
a) Authors overrate their work or b) Reviewers underrate most papers? Source: https://
blog.ml.cmu.edu/2022/11/22/neu
rips2021-author-perception-experiment/
… https://
cs.cmu.edu/~nihars/prepri
nts/NeurIPS_2021_Author_Perception_Experiment.pdf
…
@thegautamkamath
-
NeurIPS Authors Overestimate Acceptance Chances Dramatically
By
–
-
AI’s Impact on Academic Life and Higher Education
By
–
An academic's worst nightmare? Opening of "Stoner," by John Williams.
-
Graduate Research Programs in Computer Science Overview
By
–
We're talking about graduate research programs in computer science. Also, I did all my post-secondary education in the US. I know the US system pretty well.
-
University Rankings: Validation or Educational Guidance?
By
–
I do question though, whether these use cases (high schoolers and undergrads choosing which programs to apply to) are the reason that people advocate for them. Or whether it's some sort of validation for the individual, "look at how good I am/my school is"
-
Research Mentorship for Undergraduates at Less Research-Active Schools
By
–
Right. But at that point, at a good school, hopefully a research mentor should be able to guide them. Still useful for undergrads at schools which are less research-active, or for students with absent research mentors.
-
University Rankings as Educational Access Tool for Underserved Populations
By
–
Yes, I think this is the most (only?) meaningful use case for such rankings. People who may not learn these things from their standard environment. FWIW, I probably learned about my undergrad (Cornell) similarly.
-
Academic Rankings in AI Research: Beyond Simple Metrics
By
–
I actually don't mind if rankings are there as an option. Like if you want to find out who's "best" by # of ICML papers + 0.5 * # of NeurIPS papers normalized by author, or whatever weird formula, go wild. But that shouldn't be the main point.
-
CSRankings Gaps and ArXiv Collaboration Visualization Tools
By
–
Yes! CSRankings falls short of this goal, since some areas are wholly unrepresented (a student wanted to know who works on learning theory… but COLT isn't there). @RoyRinberg showed me this nice tool he built to visualize arXiv collabs
-
Why Do We Need AI Rankings Anyway?
By
–
I see a lot of people in the replies saying, but it's the best ranking we've got!! Remind me again… why do we need a ranking?
-
University Researchers Need Industry Cross-Appointments
By
–
Sure. I had a specific use case that only needed university researchers. But it was still missing some with industry cross appointments.