As an experiment for the STOC 2026 conference, we are offering authors the opportunity to receive pre-submission feedback on their papers generated by an advanced LLM-based tool based on Google’s Gemini model that has been optimized for mathematical rigor. The goal is to provide constructive suggestions to help improve your paper or help find any technical mistakes before the final STOC submission deadline. Our objective is to provide an AI tool for theoretical computer scientists, by theoretical computer scientists and identify the best way to serve the community going forward. We believe an active engagement with the community will help find the best way to empower researchers with advanced AI methods.
This experiment is conducted by PC members David Woodruff (dwoodruf@cs.cmu.edu) and Rajesh Jayaram (rkjayaram@google.com), as well as Vincent Cohen-Addad (cohenaddad@google.com) and Jon Schneider (jschnei@google.com). Please reach out directly to this organizing team, or to the STOC PC chair (aczumaj@acm.org), for any general questions or concerns regarding the process. You will be able to enter any conflicts with this organizing team through HotCRP.
This program is entirely optional and opt-in.
The reviews generated WILL NOT be passed on to the PC. The reviews are only visible to the authors and to the organizers of the program.
Authors who opt-in via the HotCRP submission form by the deadline below will receive a review generated by an advanced LLM-based tool optimized for checking mathematical rigor built on top of the Gemini model. Some examples are provided on the following web page.
Experimental Review Deadline: To be eligible for this program, at least one author of the submission must opt-in and submit the full paper by November 1, 5pm EST on the HotCRP system. If you opt-in, we will make our best effort to respond to you over email within 48 hours of receiving the current version of your paper on HotCRP, which can be any time before November 1. We may extend the November 1 deadline if absolutely necessary.
We will allow for multiple submissions to the program but will prioritize first-time submissions.
For resubmissions, you can directly respond to the email you receive from the organizers with your initial feedback.
By opting into this program, you agree to the following terms.
No Logging: We commit that your submitted paper will NOT be logged, stored, used for training the model, or used for any purpose other than generating this one-time automated review.
Access Control: Access to your paper is strictly limited to the automated pipeline. In the rare event of a technical failure, a minimal subset of the STOC organizing team above (David Woodruff, Rajesh Jayaram, Vincent Cohen-Addad, and Jon Schneider) may access the paper solely for the purpose of debugging the pipeline. This access will be in strict compliance with the conflicts you list with this organizing team in HotCRP.
Confidentiality of Review: To help us improve the tool and understand its output in context, we request that you discuss the review with someone on the organizing team - David Woodruff (dwoodruf@cs.cmu.edu), Rajesh Jayaram (rkjayaram@google.com), Vincent Cohen-Addad (cohenaddad@google.com), or Jon Schneider (jschnei@google.com) - before sharing it publicly.
Feedback: Your feedback is a key part of this experiment. We hope that you can fill out a short, confidential feedback form. This feedback will allow us to assess the experiment for roles in future theory conferences.
Please note that the model primarily assesses the paper's self-contained mathematical rigor. It does not possess external, area-specific knowledge, such as standard or folklore results. The model may incorrectly flag sections that rely on such unstated assumptions.
How to Opt-In: To participate, you will have to check the Opt-In box on the HotCRP submission form for your paper.
Any questions about the Experimental Program: Automated Pre-Submission Feedback should be directed to David Woodruff (dwoodruf@cs.cmu.edu), or other organizers Vincent Cohen-Addad (cohenaddad@google.com), Rajesh Jayaram (rkjayaram@google.com), or Jon Schneider (jschnei@google.com), or to the STOC 2026 PC chair Artur Czumaj (aczumaj@acm.org).