Paper Review
Specific Instructions for this Course:
- Each student needs to register two different papers listed in the Google spreadsheet (link) - put your name in the blank of the paper that you want to review (first register, first get). At the moment, the spreadsheet can only be viewed. Editing will be allowed from Nov. 02, 2023.
- For each lecture (starting from Lecture 2), I listed two research papers that are highly relevant to the corresponding topic.
- At most five students can register for the same paper, so you may want to put your name on the spreadsheet as early as possible.
- If we run out of research papers (i.e., the number of class attendees is more than 48), I will provide additional research papers for review, so no need to worry.
- The deadline for registering review papers is Nov. 30, 2023 - no changes are allowed to be made after this date.
- The deadline for submitting your two reviews is Feb. 08, 2024. Note that this due date applies to every review paper and student.
General Guidelines
The review should aim to address the following questions:
- What is the problem addressed by the paper?
- What was done before, and how does the paper improve prior works?
- What are the strengths and the weaknesses of the paper?
- What part of the paper was difficult to understand?
- What are possible improvements or further implications of the paper?
The review should be at most 2 pages using the NeurIPS 2023 LaTeX template. If you do not have any prior experience in reviewing a paper, then I strongly recommend you to go through the ICML 2023 Reviewer Tutorial for detailed instructions and examples on how to write good reviews.
Expectations from the Instructor
A paper review typically consists of three parts: a brief summary, detailed discussions about the pros and cons of the paper, and a list of possible directions for further improvement. However, there are no fixed standards to tell a good review from a bad review, as different people might have different opinions about a single paper, depending on their knowledge and expertise. It is okay that you have limited background knowledge about the presented topic, but you should understand the key arguments of the paper well and how the authors show evidence to support those arguments.
When writing the review, you should explain your opinions as clearly as possible. You are expected to provide insightful comments with pointers to specific arguments made by the authors or results demonstrated in the paper. For your comments to be insightful, you should provide reasons to further explain your comments. When you write down your thought process, you can better judge whether your comments are clear and convincing enough for readers to understand. As a reviewer, you are expected to be critical (not just simply agree with everything the paper shows) but provide constructive suggestions (corresponding to your critical comments) for the authors to improve the paper further.