Submission format (June 17, 2020)

[Isaac Llorente]

 Hi all,

I just made a first submission with a file named "TADPOLE_Submission_D4Live_D2_IsaacMaia-1.csv".
Upon loading this file it immediately submitted the entry, without the option to upload the "method description" document.

I suppose that either I should upload it first (before the csv), and/or that it didn't want it because it was a D4Live submission.

For clarification, as I understood, we can make live leaderboard submissions in this way, and when we are confident, we submit a file with a name like "TADPOLE_Submission_TeamName-entry".

Is this correct? Thank you in advance for your time!

-----------------------------------------------------------------------------------

PS: Not sure this is the right place to introduce myself, but I'm a master student who took this much interesting topic of Alzheimer prediction as the basis for my master's thesis. I owe a lot of knowledge to this challenge!

PS2: Also a bit off topic, but here are a couple of haikus, following the trend in the Fun section:

    I want to predict
    all the ventricle volumes;
    hopefully it helps


    This script will tell me
    the mean average error
    I hope it is small

[Razvan Marinescu]

Hi Isaac,

Welcome to TADPOLE! Glad you're enjoying the challenge.

I'm not sure if Neil or someone from London managed to fix the live leaderboard -- it broke when the owners of the master domain (grandchallenges.org) migrated away from dropbox. 

Meanwhile, you can evaluate it locally on your machine with evaluation/evalOneSubmission.py. See the evaluation/Makefile for how to run it. You will need the D4 test set from LONI, which is in the zip file.

Raz 

[Isaac Llorente]

Thank you for the quick and informative response, Raz!

While I am aware of the evaluation methods, I was hoping to have an official submission to make sure I am in the right track, and eventually for future method evaluations (with new ADNI data).

Isaac

[Neil Oxtoby]

Welcome Isaac!

Unfortunately, I haven't yet managed to fix the automatic evaluation, but you should be able to submit both your methods description and forecast CSV here:

https://tadpole.grand-challenge.org/evaluation/submissions/create/

Do you see anything like the screenshot attached?

I hope to have the evaluation system up and running soon.

Neil

[Isaac Llorente]

Yes, I already managed to submit it this way, thank you. However, just as an informative note:

I uploaded the predictions before the methods, and the page redirected automatically before I had the chance to submit the methods' description.

I will update the submission anyway with a more statistically sound way of estimating the confidence interval,

and eventually, I also want to submit approaches with 'minimalistic' input features (only blood-sample related, and/or MRI, for example, since they are less invasive than CSF sampling and also radiation-free).

- I suppose that if I give the same name it will override my previous submission; is that correct?

- I was also interested in joining the future stages of this challenge, but I suppose you will make an official announcement and set up a deadline.