AISTATS 2024 paper reviews are supposed to be out today. I thought to create a discussion thread for us to discuss any issue/complain/celebration or anything else.
There is so much noise in the reviews every year. Some good work that the authors are proud of might get a low score because of the noisy system, given that AISTATS is growing large these years. We should keep in mind that the work is still valuable no matter what the score is.
You must log in or register to comment.
A/BR/BR/BA. I feel like 2 just didn’t like the paper and 3 wasn’t familiar with the literature so I’m feeling good for the responses.
BA/BR/BA/BA/A. How good are the chances?
In the exact same boat as you. 3 BA, 1 A, 1 BR. Feels like a coin flip at this point. I tried hard to find stats on previous years to see what the average score was for accepted papers, but couldnt find any info
The strange thing in my case is that BR says mostly positive things about the paper in the review text, similarly to BA reviews. Maybe we can sway BR to BA, then the situation should be pretty good.
It’s a similar situation in my paper, the person who has the fewest comments is the BR person, who complains about two things: one is shared with the others, so I get it, and the other is complete misunderstanding and a very minor one. Yet, everyone else gave BA/A. In fact, A has more points of criticism than BR.
In situations like this, I feel like the reviewer just wants to give a rejection and is looking for an excuse. Maybe they have a submission that they think might be with the same area chair, too? I don’t know, maybe Im too jaded.
I agree that if we sway a BA to A or maybe a BR to BA, there is a chance for the paper to get accepted. Ive heard in the past people lowered their score even when there wasn’t new negative info in rebuttal. Im hoping that wont be the case. Wishing you the best of luck!
I got a A/BA/BR/R. Does it makes sense to go for a rebuttal and get the reviews flipped?
Thanks!
Slightly worse here: A/BA/R/R… contemplating if it’s worth it with too
Where is the space for rebuttal? I can only see the reviews but there’s no box to submit an answer.
You submit a 1 page PDF as the “author response” by Dec 5 AoE. The template for it is on the website on the FAQ page (also in the email). That’s it. There is no point-by-point response and as far as I understand, no “discussion” with the reviewers. Just an opportunity to clear any misunderstandings or a short experiment or two that fits into that 1 page
For European researchers, the release timing is not good for your sleep… 😅
I mean, if they release them at 27 Nov, 23:59, AoE, it would be 28 Nov, 11:59, GMT. So it depends how early they are I guess
HI all first time submitter and I can’t find this info anywhere. Will all papers get 5 reviews? E.g if you have 4 should we expect another one to drop?
All papers should have at least 3.
First time submitter. A, BA, BA. Should I expect more reviews? Also, is this sufficient for acceptance or BAs need to improve the score?
Same with you. But the confidence of my reviews are 3/1/3 (with maximum 5). The reviewer with 1 confidence emphersized his unsimiliarity in the whole review…
worst reviews ever. Some of them even didn’t read the paper. For example, a reviewer pointed out I have to cite another paper and kindly showed the title of it, which I’ve already included in the paper. And the reviewer gave me reject, with maximum confidence :)
Similar. I said I ran five individual runs and reported the average results. However the reviewer asked me to show how many random seeds I used. It’s also a reject with max confidence :)
worst reviews ever.
one reviewer told us we missed a dot at the end of an equation. reject!
Same things happened to me in NeurIPS/ ICLR : ( And they never showed up during rebuttal
I got A/A/BR/R, and the “R” reviewer put maximum confidence. Is it worth to trying rebut this? Also, I’m kind of surprised on the variance of the reviews.
You have two As. You should submit rebuttals without any doubt.
Thanks, I’ve talked to my advisor and we discussed it’s worth a try mainly because the arguments of R are rebuttable. I’ll be updating what happens at the end of the line.
BR,BR,BR,BA. Am I hopeless？
A/A/BA/BR. However the BR has maximum confidence : (
consensus seems towards an acceptance, so perhaps it will work out :)
Question for people who have reviewed before (this was my first time reviewing): It seems like reviewers don’t see other reviewers’ comments on the same paper, unlike NeurIPS. Will this change after the author response time? Otherwise how are the reviewers meant to discuss things?
Question for people who are familiar with AISTATS review. Can the reviewer change their score after the rebuttal?
What’s the acceptance score ?
I got 3 BA and 1 Reject, any chance for my work to be accepted? 😢
Depending on the content of the Reject, I’d go hard on the 3 BA to push them to A. It may happen that in the end the Reject looks like an outlier. Also, confidence scores are relevant.
Are there any resources that show score distribution from previous years?