type
Post
status
Published
date
Apr 25, 2024
slug
second-paper-review
summary
这段经历告诉了我如何成为一个好老师，如何教别人。
tags
academic
review
research
category
Academic
icon
password
This week, my advisor assigned four papers for me to review, covering both familiar and unfamiliar areas. To improve the quality, I read related literature, used ChatGPT, and discussed with classmates. Based on all that, I wrote my review comments. But my advisor's feedback caught me off guard — he said my observations were shallow, which left me confused and hurt.
I had some prior review experience (my first time), but no formal training. My advisor criticized my work without specifying what was wrong. I just did my best based on my own understanding of what a review should be. Come to think of it, criticizing a student's review is, in a way, criticizing yourself as an advisor — if students could already do it perfectly, why would they need guidance?
After talking to friends, the consensus was that the advisor's reaction seemed unreasonable — he provided no guidance but criticized the result. Of course, friends might just be trying to make me feel better.
I decided to revise my review based on my own judgment and politely ask the advisor for specific guidance on how to improve.
Despite how advanced information networks are today, I believe relying solely on internet searches and self-study for academic reviewing is unrealistic and insufficient. I chose to do a PhD not just for funding, but to solve scientific problems under an advisor's mentorship.
For a paper, if its strengths are obvious, no need to over-emphasize them; if its flaws are obvious, no need to over-criticize either.
This experience made me realize that as students, we need clear guidance and hands-on opportunities to develop real skills. I hope to learn more from this — not just to complete tasks, but for genuine understanding and innovation.
On Friday evening (4/26), I discussed with my advisor about how to evaluate academic papers. Key takeaways:
- Writing quality: Is the paper readable? Does it draw the reader in and guide them through the topic?
- Clarity of arguments: Are the points clear and professional? Does the research demonstrate its value?
- Novelty: Does the framework overlap with existing work? If so, what are the clear innovations?
- Experimental design: Is it thorough? Are the datasets convincing? Are all variables considered?
- Hyperparameter analysis: Is it included? Are the hyperparameters optimized?
These criteria can help us systematically evaluate the quality of an academic paper.
Loading...
