Mengchen is from Wafei Temple
量子位 | 公众号 QbitAI
From this year onwards, high school students will also be involved in AI top meetings, is it really a computer that starts with a doll?
NeurIPS officially solicited papers from high school students, and as soon as the news came out, netizens immediately blew up, and the reaction was like: ???。
The official announcement has been released, specifically for the call for papers in the direction of "The Social Impact of Machine Learning".
- Selected papers can be presented on the NeurIPS homepage
- Authors of up to 5 winning projects will be invited to attend the NeurIPS conference in Vancouver
- It was also specifically mentioned that authors were welcome to use large models as a research tool
就，emmm……
Overnight, it caused widespread controversy in major communities in the academic circle.
One of the more concentrated arguments of the opponents is that this exacerbates inequality in educational resources.
There are also many who feel that this will further lead to the proliferation of low-quality papers: writing papers for publication is replacing serious research.
A UCSD computer science professor even revealed that he has been inundated with emails from high school students who have applied for summer internships.
Of course, at this time, netizens who use memes to satirize are indispensable.
Some netizens who are familiar with East Asian involution culture pointed out that we all know what will happen: proxy wars.
It just feels like it's right with some other things that have been going viral recently?
It can be said that it is a kind of "dream nightmare linkage" that is not so pleasing.
NeurIPS开设高中赛道
Whether NeurIPS's attempt will turn into a "dystopia" as everyone fears, let's take a look at what the announcement says.
Each submission must be a work that has been written entirely independently by a high school student author. We want each submission to highlight the potential to use machine learning to have a positive social impact or have a positive social impact.
Applications may include, but are not limited to: agriculture, climate change, education, health care, homeless issues, hunger, food security, mental health, poverty, water quality.
Other format requirements and other unimportant ones are skipped first, and there are the following noteworthy points in the announcement:
- The deadline for submission is June 27, two and a half months before the announcement.
- Proof of high school attendance, signed by each author's high school, on school letterhead, is required.
- All supplemental materials should be created entirely by the author and include videos, demos, posters, websites, or source code.
- The review process will be conducted by anonymous reviewers and no written feedback will be provided to authors.
The conference organizers welcomed the use of large models as an aid, but also put forward two important criteria:
- Adequately describe the specific methods used in the paper, including data processing or tools for filtering, visualizing, facilitating, or running experiments or proving theorems.
- The author is responsible for the entire content of the paper, including all text and graphics, and must ensure that all text is correct and original.
In general, the organizer did take into account some factors that could lead to unfair competition and made restrictions.
For example, it must be completed entirely by high school students, and some netizens joked that the master's and doctoral students who want to ask their middle school brothers and middle school sisters to fly can die of this heart.
But will the actual implementation be distorted? Some people have also imagined some possibilities.
The CVPR approach is more worthy of reference
At present, after the announcement of the official account of NeurIPS, there has been no official response to the question.
Judging from the opinions of netizens in major communities, in addition to the previously discussed concerns about "parent boosting", the opposing side also has several concerns:
The first is that for the academic community, junk papers will be more rampant. Others gave examples of low-quality papers that had been published by specific high school students.
The second is that for AI academic conferences, the review has been overwhelmed.
In recent years, the number of papers submitted to the AI summit has been surging, and even more so after the release of ChatGPT, the number of reviewers is seriously insufficient, which has also led to a decline in the quality of review comments.
So much so that some people have suggested that instead of letting high school students submit papers, it is better to let high school students participate in the review, and the quality of the review opinions is not necessarily worse than the current one......
In addition, it is also an additional competitive pressure for high school students who participate in the program.
It may also lead to some students being too utilitarian and overly pursuing the publication of papers and neglecting all-round development.
However, there are also people on the side who support it.
For example, a Databricks research scientist volunteered to review a high school track. He believes that the biggest challenge for high school students is figuring out what they want to do in the future, and participating in academic research is a good way to get started.
Some netizens pointed out that in fact, in the past, some high school students have indeed submitted articles to top conferences such as NeurIPS and CVPR.
Now the question is, would it be better to have a separate track for them?
Finally, CVPR, also as an AI summit, is also actively reaching out to and influencing high school students, and their approach is more praised, which is more worthy of reference.
Overall, the most upvoted neutral opinion is that NeurIPS struggles to balance top academic conferences with programs that cater to high school students. The move was too radical and did not consider the possible consequences.
To be conservative, you should start by encouraging undergraduates, or having high school students start by attending workshops.
Do you have any views or suggestions on the opening of a high school track by NeurIPS?
Reference Links:
[1]https://neurips.cc/Conferences/2024/CallforHighSchoolProjects
[2]https://www.reddit.com/r/MachineLearning/comments/1c1zesl/news_neurips_2024_adds_a_new_paper_track_for_high/
[3]https://x.com/thegautamkamath/status/1778580754785550819
[4]https://x.com/NeurIPSConf/status/1778767375791239442
— END —
QbitAI · Headline number signed
Follow us and be the first to know about cutting-edge technology trends