Writing a review
How to claim assignments, write a high-quality review, and meet your deadlines.
Thank you for reviewing. Peer review is what makes a conference's program trustworthy, and the authors will read your feedback carefully.
Find your assignments
Open the Reviewer Console from the sidebar (a PC Member opens the PC Member Console, the same console under their role's name). Your assigned submissions are listed with their deadlines and current status.
If you spot a conflict on a paper you've been assigned, decline it (see below) with the reason Conflict of interest, and don't review the paper. You will not be given that paper again.
Decline a paper that is not a fit
Open the paper with Start Review and read its title, abstract and PDF first. If it is outside your expertise, a conflict, or more than you can take on, click Decline this paper under Write Review in the panel on the right. Choose a reason (Outside my expertise, Conflict of interest, Too many papers or no time, Other) and, if you like, add a note to the chairs. The authors never see it.
- 1
Click Decline paper in the window. The paper leaves My Assignments, its review task closes, and the paper's Area Chair (or the PC Chairs when it has none) is told so they can give it to someone else.
- 2
Some programs ask the chairs to approve first. Then the link reads Ask to be released, and the paper stays yours, showing Release requested, until a chair answers.
Declining is possible until you submit a review of the paper, and until the program's cutoff (by default 7 days before the review deadline; the panel shows the date). After the cutoff, the panel says when declining closed and offers Message the chairs in Webmail instead; once you submit a review, the option is gone. A saved draft is kept for the chairs' record only; it is not sent.
Write the review
- 1
Read the paper end to end before opening the review form. On the review page, the authors' TL;DR (when they gave one) sits above the Abstract; an abstract pasted from a PDF is shown with its broken lines joined. View PDF opens the paper in a new tab, and Supplementary lists the authors' extra files (code, data, audio or video). Click a file to open it in a new tab: audio, video and PDFs play or show in the browser, and other files (such as ZIP) download. Beside them, a badge shows the paper's submission category, coloured by its type (cyan for a paper, purple for an abstract, blue for a proposal), and a green Blind Agreement badge.
- 2
The review deadline is shown in red, in bold, just above Write Review (and above Submit while your review is a draft), in the program's time zone. Once you submit, or after the review period closes, the deadline stays in the panel in smaller, lighter text.
- 3
Use the review form the program set for the paper's category; papers in different categories may have different forms. Most fields are mandatory, for example summary, strengths, weaknesses, detailed comments, and a numeric score. Text boxes have no minimum or maximum length; the word count under each box, in reviews and meta reviews alike, is only for your information.
- 4
Be concrete. Cite section numbers and equations when you point out a flaw. "Section 3.2 is unclear" is more useful than "the paper is unclear."
- 5
Keep your tone professional. Authors will read your review.
- 6
Save as draft as often as you like. Submit when you are happy. You can edit a submitted review until the review deadline.
Meta-reviews and discussion
If you are an Area Chair or Senior Area Chair, you'll also see a meta-review stage. Open the discussion thread for each paper, summarize the reviewers' positions, and write a clear recommendation for the Program Chair.
In a double blind program you never see who wrote a paper: names, co-authors and original file names are left out before the paper reaches you. Other reviewers appear as Reviewer 1, Reviewer 2, numbered per paper, and your own review is titled with your number on the paper (for example Official Review by Reviewer 4). You see their reviews only once you have submitted your own and the reviews are released to the authors (the program's Reviews released to authors date, or a released decision); never their drafts or confidential comments. Messages meant for the chairs or the area chairs are not shown to you. These rules are applied by the server, so they hold on every screen. Each review and message shows who can read it (the people icon), for example Program Chairs, Area Chair, Reviewer 2 (you). Until reviews are released to the authors, a lock reads Authors from and the date; when you write a message, Who will read it under Readers says the same before you post.
Deadlines
The review window is shown on each assignment. Each paper follows the dates of its own category, so two papers in one program can have different review deadlines. Reviewing opens on the program's Review site opens date (or after the submission deadline when there is none). Write Review and Submit stay available until the review deadline, even if the program releases reviews to authors earlier. When the Program Chair turns on review reminders, you get an email on the days they chose (for example 7 days and 2 days before the deadline) listing the papers still waiting for your review. A saved draft does not count as submitted. Submit the review and the reminders stop. Late reviews are visible to the Program Chair and may be excluded from the decision.


