Turtle Recall: Conservation Challenge
'/%3e%3c/defs%3e%3cpath fill='%23fff' d='M-120-80h240V80h-240z'/%3e%3cpath d='M-120-80h240v48h-240z'/%3e%3cpath fill='%23060' d='M-120 32h240v48h-240z'/%3e%3cg id='b'%3e%3cuse xlink:href='%23a' stroke='%23000'/%3e%3cuse xlink:href='%23a' fill='%23fff'/%3e%3c/g%3e%3cuse xlink:href='%23b' transform='scale(-1 1)'/%3e%3cpath fill='%23b00' d='M-120-24v48h101c3 8 13 24 19 24s16-16 19-24h101v-48H19C16-32 6-48 0-48s-16 16-19 24z'/%3e%3cpath id='c' d='M19 24c3-8 5-16 5-24s-2-16-5-24c-3 8-5 16-5 24s2 16 5 24'/%3e%3cuse xlink:href='%23c' transform='scale(-1 1)'/%3e%3cg fill='%23fff'%3e%3cellipse rx='4' ry='6'/%3e%3cpath id='d' d='M1 5.85s4 8 4 21-4 21-4 21z'/%3e%3cuse xlink:href='%23d' transform='scale(-1)'/%3e%3cuse xlink:href='%23d' transform='scale(-1 1)'/%3e%3cuse xlink:href='%23d' transform='scale(1 -1)'/%3e%3c/g%3e%3c/svg%3e)
Helping Kenya
$10 000 USD
Completed (almost 4 years ago)
Classification
Computer Vision
753 joined
247 active
Start
Nov 19, 21
Close
Apr 21, 22
Reveal
Apr 21, 22
'%3e%3cpath fill='%23001489' d='M0 0v6h9V0z'/%3e%3cpath fill='%23e03c31' d='M0 0v3h9V0z'/%3e%3cg stroke='%23fff' stroke-width='2'%3e%3cpath id='d' d='m0 0 4.5 3L0 6m4.5-3H9'/%3e%3cuse xlink:href='%23b' stroke='%23ffb81c' clip-path='url(%23c)'/%3e%3c/g%3e%3cuse xlink:href='%23d' fill='none' stroke='%23007749' stroke-width='1.2'/%3e%3c/g%3e%3c/svg%3e){kind=small}
Some motivation
Data · 21 Feb 2022, 12:04 · edited 4 minutes later · 12
Hi all,
We know this is a tricky competition and there is some concern about the error metric.
I have made a submission that has the correct turtle in at least one of the 5 columns and received a score of 0.4.
Here is a screenshot of me submitting the reference file getting a perfect score. It is possible ;)
Before the end of the competition, I will remove my submissions.
Keep going and let's see how close you can get!
Wow! The gap is no joke!..........hang tight, we're coming....lol.
Lol... it's really no joke my brother.
Surprising score.......
What did you do? Use Super machine learning model ? Or some magic ??
I work at Zindi (sorry it isn't on my profile) so I have access to the backend.
There is no ml in that score, just a copy-paste and a shuffle from the ref file ;)
I just wanted to show that the scoring metric is correct, this is just a really difficult challenge.
I will remove my submissions before the challenge closes so it is still fair.
Thank you for quick explanation and sorry for my misunderstanding ！
I'll continue to try this difficult task : )
:)
Smiles,,
Hi @amyflorida626,
Thanks for taking a look to it.
Which percentage of images are used for the public leaderboard?
I am trying to understand why everyone is getting such a good local validation... and bad lb score. Are we sure that the images are correctly tagged in the backend? (I dont find other explanation XD)
Wow Amy, you know your turtles!
Lets shellebrate.
Thanks for letting us know that the score can indeed be that high. Based on also the post by @Johnowhitaker it seems people are starting to use better models. It could make sense to perhaps do another post to share what ideas are working better for this task.
@amyflorida626
@fnoa and others have raised pretty critical analysis of the discrepancies between local cv score and the lb score.
While the challenge remains difficult could you assure us that:
1) the reference file you use is correct in the first place, try looking manually at it manually. Even a small sample of it should be perfect. It was made by humans and could have errors in terms of the turtles actual identities;
2) their is no bug in the script scoring our submissions.
These two I think would be the least assurance I would ask, and I guess on behalf of many
@amyflorida626 Thanks for looking at the reference file. Now, we code!!