CSPaper: review sidekick

Lol! this is a great idea

Yeah. Can't wait to see how AAAI 2026 First AI-Assisted Peer Review performs.

What?! Dataverse picking this moment to bail. Totally feeling your pain.

I also heard some negative ones. [image: 1747115726301-9ec54026-13df-43c4-90cf-d0c1b5e651f2-image.png] [image: 1747115784562-df6bcec1-bf5b-446d-a626-2037490ea36f-image.png] but it seems iccv doesn't allow the usage of chatgpt etc in review process: [image: 1747115959677-6e4057c2-55e9-45c7-b1e7-4b6ff2645928-image.png] [image: 1747115805501-2b07b416-a61c-48b9-9039-c45b7322dbe3-image.png] [image: 1747115830768-cd7ee90a-e148-4a88-9045-ed712cf087d3-image.png] Let's hear the reviewer side of story, but didn't say why: [image: 1747116023274-8365cfc3-cceb-4c42-9b81-27623bd26a7e-image.png]

ICLR 2026 will be hosted in Brazil. [image: 1748105858945-710873cc-02f7-4cd3-a955-4146af74d285-image.png] What's this one in Portugal? [image: 1748105973493-c47e8394-99cf-4e5d-bdae-022a01292d2f-image.png]

Been there, felt that. Sometimes peer review feels more like roulette than rigor; but hey, good science endures beyond a stray reviewer’s “hot take.”

This year, the effective number of submissions of IJCNN is around 5,000+. “All Neural Network roads lead to Rome” — this year’s IJCNN slogan is really clever. From the early stages of submission, the PC (Program Committee) had already been motivating ACs, anticipating that the IJCNN acceptance rate would align with that of top conferences. For those whose supervisors don’t reimburse for “CCF-C” (a conference ranking system widely adopted in China) conferences, you might have some questions about the registration fee. From what I know, last year one registration ($915) could cover two papers (provided the same author is registering), and additional charges only started from the third paper.

I believe this is not the only case, have seen more of alike.

@Sylvia got a similar experience from my COLING 2025 submission. A competitive track that I submitted my work to ... Meta review said literally "ACCEPT", yet still rejected as the final decision

@root I am not sure how much "push" this communication email to reviewers will do ...

The early bird deadline is June 18th! Register on or before the deadline to receive discounted rates for KDD 2025!

For authors who resubmitted their work to NeurIPS: Although opt in making the ICML rejection public can drive transparency and demostrating improvement. However, you need to think about the Cons too. There’s research (https://arxiv.org/pdf/2011.14646) showing reviewers tend to score resubmissions lower if they know it got rejected before, even if it’s the exact same paper. Plus, having the rejection out there might give the impression the paper’s not great quality, consciously or subconsciously affecting reviewers.

2 rejects [image: 1746164634117-screenshot-2025-05-02-at-07.43.22.png]