Dear FeTS Challenge Participants, We have received requests from participants for the extension of submission deadline for Task 1 of FeTS2022 challenge. Considering this, We have now extended the submission deadline to July 15, 11.59 PM EST (no further extension). Please download the official template for short paper from here: The manuscript is expected to be 10-12 pages and include the following sections: Abstract Keywords Introduction Methods Results Discussion Acknowledgments/Funding (if any) References. CMT link for the submission of short paper: ( Please select appropriate track when you create the submission. Code/Weight submission for Task 1: Once you create a submission on CMT, please write to with your team name and the CMT submission ID. Upon receiving this mandatory infomation, a unique upload link will be shared with you to submit the code. Make sure you create the submission well in advance to receive the submission link. You will be allowed to edit the submission on CMT till July 15 11.59 PM EST . Thank you for your interest in this year's challenge, and we look forward to your contributions! Feel free to share your doubts in the discussion forum. Regards, Ujjwal Baid (on behalf of the organizing committee)

Created by Ujjwal Baid ujjwalbaid
May I ask when we will receive the link to upload the code? We have sent the email earlier. Thanks.
When you run the code submitted, to control the randomness, will you apply the same seed to every submitted code? or we need specify the seed in the code ourselves? Thanks.
Hi @Johannes_Pae, As mentioned above please write to with your team name and the CMT submission ID.
We would also kindly ask for an upload link for the code. Also, we should anonymize the short paper submissions, right ? :)
Hi @frank_wang35, We have emailed you the submission link to upload the code. We will run the code once and then take the resulting model.
@ujjwalbaid Dear organizer, we have submitted the short paper, and send email to apply for the upload link of code/weight. Awaiting your reply as of now. One more question we are not so sure about: in the next evaluation phase, how will you re-train the model using our submitted code? Will you try multiple times and pick the best model you will obtained from these trials? Or you just run our code once then take the resulting model? Thanks so much, looking forward to your clarifications.
@ujjwalbaid one more quick question: during training, every one of the selcted collaborators runs validation once. How big is this validation set in comparison to the training set? Rough estimate is fine, I assume something like 20%?
Ok super thank you for the quick help!
You just need to submit the code.
Thank you for the quick answer! What if I dont have a best model though? I assumed I dont have to spend weeks on training and just submit the algorithm.
Hi @anon123, It refers to the best model which is saved.

Updates: Deadline extension for Task1, CMT Submission System for short papers, Submission of Codes and Weights (Task 1) page is loading…