Primary competition visual

ICLR Workshop Challenge #2: Radiant Earth Computer Vision for Crop Detection from Satellite Imagery

Helping Kenya
$5 000 USD
Completed (almost 6 years ago)
Classification
Earth Observation
651 joined
110 active
Starti
Feb 03, 20
Closei
Mar 28, 20
Reveali
Mar 29, 20
User avatar
ZINDI
Final Leaderboard
Connect · 16 Apr 2020, 14:03 · 3

Dear Zindians,

Congratulations to the winners of this challenge!

Thank you for your patience while we reviewed the winning solutions. You might notice that there has been a major shake-up on the leaderboard.

Some winners did not submit their code for review and some users used the restricted features, so we have re-run their solutions without these features to give them the new score.

We have shuffled users to the appropriate position and have matched the score we generated from code review to the closest score users have submitted. Thank you for alerting us to the data leaks, and for your patience while we reviewed and ranked your submissions.

Please note that we have increased the code review process to 3 weeks for future challenges as we are now reviewing more solutions.

To learn more about our updated rules, see this blog post.

For tips on code submission, see this blog post.

Discussion 3 answers

Dear Zindi team,

I have one question. Is it possible to have access to the leaderboard test data? I would like to use your dateset to explore further solutions and approaches and link these results to some other internal similar datasets. If not possible, can you open a new leaderboard where researches can test their approaches?

Hope to hearing from you soon.

Dario

20 Apr 2020, 13:56
Upvotes 0
User avatar
ZINDI

Hi Dario,

Unfortunately, we cannot share the reference sets however you can look on Radiant Earth's ML Hub for similar data sets to practice one.

You can still submit to this challenge and receive a score on 100% of the reference file. It will show under the 'Submissions' tab.

Thank you so much, the possibility to submit is already enough. However, I'll take a look on Radiant Earth's ML Hub as well, thank for this useful info!