Frequently asked questions
Q1: Is remote participation allowed?
Response: Yes, we will support remote participation. However, we encourage in-person attendance for oral and spotlight presentations, as it tends to foster more dynamic discussions.
Q2: Is registration necessary to participate in the workshop?
Response: If you plan to attend the workshop remotely, you must register for remote ICML workshop attendance. Details can be found at https://icml.cc. If you are only submitting a paper and do not plan to attend, registration is not required. You'll have the option to post your paper and notebook on the website even if you are not registered.
Q3: Can I contribute to the workshop as a reviewer?
Response: Thank you for your interest! Please send us an email at icml2025-moss-workshop@googlegroups.com with a brief description of your experience and/or Google Scholar profile. To ensure the review quality, we require reviewers to have published at least 2 papers.
Q4: What should the notebook contain?
Response: The notebook should provide convincing evidence that the main claims of the paper are correct. There can be additional results in the paper but reviewers should primarily judge the work based on the claims which can be reproduced by running the notebook. You do not need to replicate all experiments; just enough to convincingly validate the important claims.
Q5: Can I submit theory papers?
Response: Yes, theory papers are welcome! It’s encouraged that notebooks include experiments that either validate the paper’s assumptions, demonstrate that the theoretical claims hold more broadly, or illustrate the results in a compelling way.
Q6: Can we submit works that require API calls?
Response: Only free APIs! No paid OpenAI API use. This ensures that all submissions remain reproducible and accessible to reviewers and other researchers.
Q7: Does the entire codebase need to be in the notebook?
Response: No! You can submit your entire codebase as a zip file and a notebook can simply import functions from your code files. The goal is for people to be able to reproduce your results and build on them. Reviewers will be asked to check the reproducibility of the main notebook, and that key supporting experiments should be included there.