Link: https://www.zhihu.com/question/597314456
Editor: Deep Learning and Computer Vision
Disclaimer: For academic sharing only, intrusion and deletion
Author: Zhihu users
https://www.zhihu.com/question/597314456/answer/3000946712
Not because the article was rejected, but because of the most outrageous AC that refreshed my three views so far. 
Fact: This work is the most discussed and most important baseline in our article
I don’t mention that there are only 2 reviewers in this article, and I don’t mention that one of the reviewers can’t tell the difference between ODE and SDE. I rejected the sentence “ODE is not flexible enough” (the word ODE doesn’t even appear in the full text), and I said that in the end AC gave Reasons for refusal-important omission: Did not mention/discuss another related work-in fact, this work is one of our most important baselines, clearly placed in all experimental links and diagrams of the paper, Because the effect is very good, we also spent a considerable amount of time analyzing its reasons and mechanisms. I can't find a suitable reviewer, I can't bear it, I can't read the paper, I can't bear it, I can't read it carefully, so righteously confusing black and white, making something out of nothing, pointing a deer as a horse, and finally humiliating you with a condescending yin and yang: "I'm surprised The authors didn't even discover the prior work!" I can't fucking stand this! Ironically, the sentence in the email I received for another article that was published: Each submission was reviewed by area chair and senior area chair, in order to ensure each submission is asseseed properly—— so "asseseed properly" is noble The AC can talk nonsense with his eyes open and open his mouth during the sharp review, and he doesn't even bother to use ctrl+F. What's more ironic is that the prior work he mentioned was "missed" by our group last year.
Author: Jiapeng Zhang
https://www.zhihu.com/question/597314456/answer/2999435881
Originally, the paper was going to vote for COLT. As a result, because ICML changed the conference address to Hawaii, two of them were resubmitted to ICML with the mentality of traveling. It turned out to be the dumbest decision I made this year. Submitting articles to ICML is a waste of time. In the future, all articles prepared in January and February will be submitted to COLT.
The quality of COLT's review is significantly higher. Both the advantages and disadvantages of the article can be accurately grasped. ICML is basically playing the piano against the cow. I feel that the reviewers and AC basically didn't read our rebuttal.
作者：Hunto
https://www.zhihu.com/question/597314456/answer/3007065905
I am curious why this year's ICML decision is not visible to the reviewers. There are a few controversial papers in the review, and I still want to see the results.
In addition, I would like to complain about the quality of ICML's review. In my field, there are many reviewers who just fool around and write clichés. During the discussion stage, I pretended to be dead and no one discussed it, and no one responded to the discussion I initiated. In comparison, the reviewers of NeurIPS and ICLR are more active.
Author: Da Xixi
https://www.zhihu.com/question/597314456/answer/3000609636
A meta review with only one sentence, the most disappointing review ever received, the key is that the rebuttal reviewers didn’t reply anything: After the rebuttal phase, the reviewers are still not convinced that their comments were fully addressed and clarified by the authors. Hence there is A common consensus that there is still room to further improve this paper, both in presentation and technical improvements.
作者：Hokjau
https://www.zhihu.com/question/597314456/answer/2999878752
The acceptance results of ICML 2023 reflect the fierce competition in the global computer science field. According to reports, the number of applicants exceeded 25,000, and the acceptance rate was only 1.2%. This shows that talent in the field of computer science is very scarce, and everyone has a chance to get top opportunities in this field.
The recruitment results of ICML 2023 also reflect the strict screening of candidates by recruitment agencies. The admission results show that the applicants have excelled in multiple areas, including algorithmic ability, research results, programming skills, teamwork, etc. This shows that ICML 2023 recruitment agencies pay great attention to the overall quality of candidates, not just a certain aspect of skills.
The acceptance results of ICML 2023 also illustrate that the field of artificial intelligence and computer science is developing very fast. With the continuous advancement of technology and the continuous expansion of application scenarios, the demand for high-quality computer science talents is also increasing. Therefore, people working in this field need to keep learning and upgrading their skills to adapt to the ever-changing market demands.
The admission results of ICML 2023 show that the competition in the field of computer science is fierce, and the requirements for comprehensive quality in this field are also very high. People working in this field need to keep learning and upgrading their skills to adapt to the ever-changing market demands. At the same time, the acceptance results of ICML 2023 also reflect the rapid development of this field and the wide application of artificial intelligence technology.
☆ END ☆
If you see this, it means you like this article, please forward and like it. Search "uncle_pn" on WeChat, welcome to add the editor's WeChat "woshicver", and update a high-quality blog post in the circle of friends every day.
↓ Scan the QR code to add editor↓
