Hello everyone,
Again thank you all for your participation in DeftEval! The official subtask 1 and 2 leaderboards are now available on Github and on the front page of the Codalab site. Because we heard from many of you who did not submit your best score to the leaderboard,
or forgot to include your other subtask results in your leaderboard submission, we made the decision to use the
best result from each user regardless of whether or not you submitted that result to the leaderboard. If for whatever reason (interesting architecture, alternative method, etc.) you intentionally submitted a different result to the leaderboard, remember
that you are still absolutely welcome to submit those ideas to the conference regardless of your leaderboard submission.
Additionally, if you placed lower than expected in the leaderboard, remember that you are still welcome to make a paper submission to the SemEval conference. We are interested both in high performing models and models that are interesting in architecture, approach, or otherwise significant even though they may score below the top 5, top 10, etc.
If you have any questions, please feel free to reach out. We will post the results of subtask 3 as soon as we have the aggregate data together after the submission deadline tomorrow.
Best,
Sasha
From: <semeval-202...@googlegroups.com> on behalf of "hx...@georgetown.edu" <hx...@georgetown.edu>
Date: Tuesday, March 10, 2020 at 11:37 AM
To: DeftEval 2020 <semeval-202...@googlegroups.com>
Subject: Re: Subtask 1 and 2 Results
Hi Sasha,
Thanks for posting the results. I'm just wondering how is this result calculated? Seems like it's not purely F1 score.
Kind regards,
Haotian Xue
On Tuesday, March 10, 2020 at 11:27:35 AM UTC-4, Sasha Spala wrote:
Hello everyone,
Again thank you all for your participation in DeftEval! The official subtask 1 and 2 leaderboards are now available on Github and on the front page of the Codalab site. Because we heard from many of you who did not submit your best score to the leaderboard, or forgot to include your other subtask results in your leaderboard submission, we made the decision to use the best result from each user regardless of whether or not you submitted that result to the leaderboard. If for whatever reason (interesting architecture, alternative method, etc.) you intentionally submitted a different result to the leaderboard, remember that you are still absolutely welcome to submit those ideas to the conference regardless of your leaderboard submission.
Additionally, if you placed lower than expected in the leaderboard, remember that you are still welcome to make a paper submission to the SemEval conference. We are interested both in high performing models and models that are interesting in architecture, approach, or otherwise significant even though they may score below the top 5, top 10, etc.
If you have any questions, please feel free to reach out. We will post the results of subtask 3 as soon as we have the aggregate data together after the submission deadline tomorrow.
Best,Sasha
--
You received this message because you are subscribed to the Google Groups "DeftEval 2020" group.
To unsubscribe from this group and stop receiving emails from it, send an email to
semeval-2020-task...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/semeval-2020-task-6-all/e1927cc1-f0a9-4ab0-a7df-bc7e5b05b2bf%40googlegroups.com.
To unsubscribe from this group and stop receiving emails from it, send an email to semeval-2020-task-6-all+unsub...@googlegroups.com.
--
You received this message because you are subscribed to the Google Groups "DeftEval 2020" group.
To unsubscribe from this group and stop receiving emails from it, send an email to semeval-2020-task...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/semeval-2020-task-6-all/7E51B244-173C-4B08-B552-F0B3689FB64A%40adobe.com.