Click the Card below and follow the “CVer” public account
AI/CV heavy-duty information, delivered as soon as possible
Click to enter->[Target Detection and Transformer] Communication Group
Reprinted from: Xinzhiyuan | Editor: Editorial Department
[Introduction]An excellent ACL paper was rejected because it was uploaded to arXiv half an hour after the anonymity period? ACL has attracted a lot of criticism!
ACL, causing public outrage!
This morning, the entire AI community was in an uproar because of this post–
Naomi Saphra, a postdoctoral fellow at New York University, said that she had just received a direct rejection from ACL because the paper (signed) was still submitted to arXiv after the anonymity deadline, only 30 minutes late.
Naomi Saphra angrily denounced: ACL’s embargo policy has seriously harmed junior researchers and made the ACL conference no longer suitable for NLP work.
Machine learning guru Sebastian Raschka also expressed support–
Note that this paper was not submitted to the conference a few minutes late, just a few minutes late to an unnecessary external preprint server.
On Twitter, angry researchers gathered under the post and launched a crusade against the ACL competition organizing committee.
This event
Saphra was rejected because ACL had such a policy.
After 11:59PM (UTC-12:00) on December 20, the anonymity period begins, and papers submitted to ACL may not be published in non-anonymous form thereafter.
It is particularly worth noting that for preprint papers submitted to arXiv, the ACL regulations refer to “submission time” rather than “publication time”.
Official website address: https://2023.aclweb.org/calls/main_conference/
In addition, ACL does not prohibit researchers from speaking on the work under review, but during the anonymity period, researchers are not allowed to promote preprints or speeches on social media, talk about them in blogs, or let the media report them.
After the anonymity period has begun, anonymous preprints can be published, but the authors or their close colleagues cannot publicize them, as this may affect the review process.
Another article, also 25 minutes late
Interestingly, MIT undergraduate Zack Ankner, who co-authored another paper with Saphra, was also rejected because the paper appeared on arXiv 25 minutes after the anonymity period began.
Paper address: https://arxiv.org/abs/2305.15096
The start time of the EMNLP anonymity period is as follows:
Ankner said that he was very confident in the review results and spent a whole week writing rebuttal and coding experiments.
As a result, the paper was still rejected.
As the second author of the paper, Saphra said, “This is the first paper submitted by Zack Ankner. I told him that the EMNLP conference should be better than the ML conference because the former has the option of short papers. But now, after experiencing so much Too bad, I don’t think the ACL conference is worth attending. More and more people think so.”
In this regard, Mosaic ML and Databricks scientist Jonathan Frankle said, “I am very angry. I will never submit to ACL again! They used stupid policies to harm a student after rebuttal. However, now that the anonymity period has expired, we Now we can promote it.”
Machine learning guru Sebastian was also shocked: “In the past, they just didn’t let authors talk about their arXiv papers online, but now they even have restrictions on when they can upload them and when they can’t?”
Netizens question
Under this post, netizens expressed doubts.
“Despite the express prohibition, I don’t understand why you were not notified at the beginning of the review period? 3 reviewers, 1 regional chairperson, and no one saw it?”
“I also had this experience at the EACL conference in 2016. The direct rejection was issued after rebuttal. If it must be rejected, why not reject it in the first week and make the contributors wait for several months?”
“Another strange thing is that the anonymity period is one month, not 30 days. If it is the next month, your paper may be fine…”
“I’ve said before that arXiv has done more to advance science and expand scholarly participation than anonymity has ever done. Policies that hinder arXiv’s mission are not only foolish, but counterproductive to scientific progress and inclusivity.”
At the same time, netizens also recalled their experiences of being rejected by the organizing committee due to various procedural issues.
“This reminds me of how my paper was rejected because I used the ACL 2021 template instead of the new ACL ARR template released a few months ago.
Rejecting the manuscript is a foolish decision because this problem can be easily solved after the paper is returned. But the thesis itself took me 3-4 months to complete. “
“We were rejected because we optimized character spacing. It makes no sense.”
“One of my papers was rejected because the student added the table title at the top of the table instead of the bottom.”
In the face of all kinds of chaos, netizens ridiculed: “ACL has long lost its scientific backbone, replaced by increasingly complex bureaucratic rule.”
Goodbye, ACL
Perhaps because they were disappointed with the actions of the ACL Organizing Committee, many netizens expressed their determination not to submit articles to ACL again.
Fu Yao, a doctoral student at the University of Edinburgh, said, “After comparing ACL 2023 and ICML 23, I asked my colleagues whether they would submit their papers to *CL, and they said no. There are many reasons. In short, large model scholars I will no longer participate in ACL and ICML.”
“Now, there is one more reason.”
CVPR ban: Prohibited publicity during the review period
In fact, this is not the first time that a similar ban has appeared.
In 2021, CVPR has stipulated that during the review period, the authors of the paper are prohibited from discussing or self-promoting in any form on social media.
The CVPR 2021 PAMI-TC meeting has previously passed a motion that paper authors should not use social media to spread papers submitted to CVPR during the review period. Social media silence period defined in motion. The social media silence period begins four weeks before the paper submission deadline and ends when notification of the final result of the paper is sent to the author. According to the current schedule, this silent period runs from October 19, 2021, to March 2, 2022. During this period, any author-initiated social media promotion of the paper will be considered a policy violation.
CVPR believes that this policy helps maintain the credibility of double-blind reviews.
This motion believes that the current policy cannot effectively restrict the behavior of submitting authors to promote their papers on social media, and the definition is vague. As long as the author does not clearly indicate that the article is in the review stage of the conference, promotion on social media is allowed.
The intensification of this kind of publicity will undoubtedly affect the reliability of double-blind review and affect the reviewer’s judgment of the quality of the article.
At the same time, influencers with many followers on social media will gain a greater advantage in public opinion, which will reduce public trust in the fairness of the review system and affect mutual trust within the scientific community.
As soon as this ban came out, the community was in a panic and there were many opponents.
Even deep learning pioneer and Turing Award winner Yann LeCun called it “crazy”!
“This policy not only limits the effective exchange and communication of scientific information, but also hinders progress.”
Everyone believes that the original intention of this policy is good, but it is difficult to implement it in practice because there are too many ways to bypass “active propaganda”.
It seems to promote fairness, but in fact it further widens the “gap between rich and poor”.
The author is famous himself and backed by a well-known institution, so there is no need to promote through social media. But for ordinary authors, if social media is no longer allowed, there is really no other way.
References:
Just got a desk reject, post-rebuttals, for a paper being submitted to arxiv <30 min late for the anonymity deadline. I talk about how the ACL embargo policy hurts junior researchers and makes ACL venues less desirable for NLP work. I don’t talk about the pointless NOISE it adds.
— Naomi Saphra (@nsaphra) September 4, 2023
Click to enter-> [Target Detection and Transformer] Communication Group strong>
ICCV/CVPR 2023 paper and code download
Backend reply: CVPR2023, you can download the CVPR 2023 papers and code open source paper collection
Backend reply: ICCV2023, you can download the collection of ICCV 2023 papers and code open source papers
Target detection and Transformer exchange group established Scan the QR code below, or add WeChat: CVer333 to add CVer Assistant WeChat, and then apply to join the CVer-Target Detection or Transformer WeChat communication group. In addition, other vertical directions have been covered: target detection, image segmentation, target tracking, face detection & recognition, OCR, pose estimation, super-resolution, SLAM, medical imaging, Re-ID, GAN, NAS, depth estimation, automatic Driving, reinforcement learning, lane detection, model pruning & compression, denoising, fog removal, rain removal, style transfer, remote sensing images, behavior recognition, video understanding, image fusion, image retrieval, paper submission & communication , PyTorch, TensorFlow and Transformer, NeRF, etc. Be sure to note: Research direction + location + school/company + nickname (such as target detection or Transformer + Shanghai + hand in + Kaka). Note according to the format to be approved and invited to the group faster. ▲Scan the QR code or add WeChat ID: CVer333 to join the communication group CVer Computer Vision (Knowledge Planet) is here! If you want to learn about the latest, fastest and best CV/DL/AI paper express delivery, high-quality practical projects, AI industry cutting-edge, and learning tutorials from entry to mastery, please scan the QR code below and join CVer Computer Vision, which has gathered thousands of people! ▲Scan the code to enter the planet ▲Click on the card above to follow the CVer public account
It’s not easy to organize, please like and watch