Patterns for writing good rebuttals（借用"我就说我对这篇论文很自豪，与是否能被录用已经没有关系了"）

作为第一次写论文就投顶会的小菜鸟，真的是有好多地方要虚心学习。焦急又兴奋的收到review的评论，读完就觉得自己没戏了，rebuttal本想随意凑合一下，但是两位老师还是拉着我一字一句地改了，指导我什么是重点，该怎么表达，觉得自己像一个学写英语作文的小学生，哈哈哈哈，虽然结果理想的可能性并不大，但是依旧感谢两位老师的陪伴，吸取经验教训。

Understand the decision process.

Most SE conferences apply a process called Identify the Champion. Carefully read these process patterns, understanding the roles of the reviewers and the PC chair, and where your rebuttal can make a difference.

 

Identify the undecided.

These are reviewers who may be swayed by your arguments. You can identify them (a) by generally liking your paper, but not the details, (b) because they provide their scores, or (c) because they explicitly state that they may be swayed. Focus your arguments on these reviewers. (And if you ever become a PC chair: include scores in the reviews, such that your authors know whom to focus upon.)

 

Identify the champion.

The champion likes your paper anyway, so don't bother too much with her review. If there's no potential champion, chances of getting your paper accepted are slim.

 

Arm the champion.

In the final discussion, your champion will need arguments against the detractors, especially strong detractors. Refute every single issue raised by the detractors to give your champion extra arguments for acceptance. Lower the confidence in the detractors' reviews by pointing out mistakes.

 

Identify the detractors.

A strong detractor can only be countered by a strong champion. Rather than trying to dissuade a strong detractor, your aim should be on arming the champion (see above).

 

Answer the questions.

Some conferences want you to only answer the specific questions the reviewers have asked, and/or instruct their PC members to only focus on these answers. Try to relate every argument of yours to a specific question.

 

Write for the PC chair.

If your reviewers have done a bad job, say so in the beginning, and the PC chair may assign you another reviewer. Stick to facts (review is just one paragraph, reviewer claims "no Foo" when Section 3 is named "All about Foo", etc.) that the PC chair can check in less than a minute. If a reviewer completely misunderstood the paper, feel free to report this, but keep in mind it is your duty to ensure even a casual reader gets the message. If all reviewers misunderstood your paper, the problem is not the reviewers.

 

Write for the committee.

When your paper is discussed during the PC meeting, reviewers will glance through the reviews, and also read your rebuttal. If there's one thing you want the committee to know, say so in the very beginning (because that's all most reviewers will read). This assumes that your paper will be discussed, so focus on getting it discussed first.

 

Convince.

Convince the reviewers that you'll be able to improve the paper. If the reviewer asks: "Do you have a formal definition of Foo?", don't say "Yes". But don't say "Thanks! We'll add this." either. Say "A Foo is a Bar within a Qux range; we'll add a detailed definition." You don't need to provide the full change; just sketch the change you will make.

 

Choose comments wisely.

Focus on major concerns - the ones where the reviewer assumes you may not be able to improve the paper. You do not need to convince the reviewer that you're able to fix typos, straight-forward presentation issues, language issues, or anything else that can be fixed by simple proofreading. This is taken for granted.

 

Organize your rebuttal.

Make it easy for reviewers to find their questions (and your arguments). Refer to reviewers by R1, R2, ...; questions by R1Q1, R1Q2, ... Use bullet points and short sentences to allow for quick scanning. Start with the most important and most common concerns, going down towards lesser issues.

 

No tricks.

If the limit is 500 words, do_not_trick_the_process by_replacing_spaces_with_underscores or likewise; all that will happen is that the PC chair will delete your rebuttal. If it's 2500 characters, do not assume that reviewers KAAYAYA and AYP because they FSITPOG (know all about you and your acronyms and accept your paper because they feel stupid in the presence of genius).

 

Thank the reviewers.

A rebuttal is not a place to vent off your fumes, even though your work is so great and the reviewers are all so stupid. Reviewing is hard work and even the least significant review contains grains of truth which will be helpful for your future work.

 

Don't expect too much.

Your rebuttal will not save your paper. Your rebuttal will play a role only if your paper is on the edge, or if reviewers made very obvious mistakes. If the reviewers update their reviews to reply to your arguments, that's already a lot. But regardless of the outcome, your rebuttal will help to improve your paper and eventually get it accepted - if not now, there's always another deadline on the horizon.

 

 

from https://www.st.cs.uni-saarland.de/zeller/onresearch/rebuttal-patterns.php3

感谢Andreas Zeller 这位德国大牛的文章:怎样写好rebuttal，反对你的审稿人。

附上LeCun的经典rebuttal,希望下次我也有这样的底气

 

Hi Serge, We decided to withdraw our paper #[ID no.] from CVPR "[Paper Title]" by [Author Name] et al. We posted it on ArXiv: http://arxiv/ [Paper ID] . We are withdrawing it for three reasons: 1) the scores are so low, and the reviews so ridiculous, that I don't know how to begin writing a rebuttal without insulting the reviewers; 2) we prefer to submit the paper to ICML where it might be better received; 3) with all the fuss I made, leaving the paper in would have looked like I might have tried to bully the program committee into giving it special treatment. Getting papers about feature learning accepted at vision conference has always been a struggle, and I've had more than my share of bad reviews over the years. Thankfully, quite a few of my papers were rescued by area chairs. This time though, the reviewers were particularly clueless, or negatively biased, or both. I was very sure that this paper was going to get good reviews because: 1) it has two simple and generally applicable ideas for segmentation ("purity tree" and "optimal cover"); 2) it uses no hand-crafted features (it's all learned all the way through. Incredibly, this was seen as a negative point by the reviewers!); 3) it beats all published results on 3 standard datasets for scene parsing; 4) it's an order of magnitude faster than the competing methods. If that is not enough to get good reviews, I just don't know what is. So, I'm giving up on submitting to computer vision conferences altogether. CV reviewers are just too likely to be clueless or hostile towards our brand of methods. Submitting our papers is just a waste of everyone's time (and incredibly demoralizing to my lab members) I might come back in a few years, if at least two things change: - Enough people in CV become interested in feature learning that the probability of getting a non-clueless and non-hostile reviewer is more than 50% (hopefully [Computer Vision Researcher]'s tutorial on the topic at CVPR will have some positive effect). - CV conference proceedings become open access. We intent to resubmit the paper to ICML, where we hope that it will fall in the hands of more informed and less negatively biased reviewers (not that ML reviewers are generally more informed or less biased, but they are just more informed about our kind of stuff). Regardless, I actually have a keynote talk at [Machine Learning Conference], where I'll be talking about the results in this paper. Be assured that I am not blaming any of this on you as the CVPR program chair. I know you are doing your best within the traditional framework of CVPR. I may also submit again to CV conferences if the reviewing process is fundamentally reformed so that papers are published before they get reviewed. You are welcome to forward this message to whoever you want. I hope to see you at NIPS or ICML. Cheers, -- [Author]