The day is finally here! 😁

How to participate in today's Catchathon?

Simply head to Stall Catchers any time between 17:00GMT (1pm ET) today (April 28) to 17:00GMT (1pm ET) tomorrow (April 29), and CATCH!

There will be double points during the final hour! So if you want to grab a bigger loot, don't forget to join us from 16:00GMT (12pm ET) to 17:00GMT (1pm ET) tomorrow (April 29)

The individual catcher and team leaderboards will be displayed on the main interface, and we will be posting regular updates here too, so don't forget to check back!

If you're an individual catcher, don't forget to choose the team you want to play for. Find the instructions for doing that here.

If you are running a school, library or another type of team, here are some resources to help you on the day:

If you need any other information or support, let us know via email, and we will do our best to help you and your team out! :) Β 

Please join our live Zoom hangouts!

We will have a kick-off hangout with our scientific team at 1pm ET today, which will be streamed live on YouTube. You and your students will be able to ask questions and we will answer them live!

And, to close the event, we are holding the final hour hangout (12pm ET on Thursday) with all the teams who want to join - so you and your team could join as panelists as well and chat with us! If you'd like to be a part of the hangout, please let us know, and we will send you the Zoom invite. Β 

What's special in this year's Catchathon?

Well, I'm glad you asked! 😜

We are going to have the first ever catcher-bot - GAIA πŸ€– - join the competition. GAIA πŸ€– is an automated catcher, who learned all her β€œtricks” from you - our human catchers! During this catchathon, GAIA πŸ€– will be catching alongside all of you. The goal is to see how well GAIA πŸ€– and human catchers can work together, to speed up data analysis even further! Look for more information about GAIA πŸ€– here on the blog soon.

GAIA πŸ€– Β was "born" out one of the winning machine learning models in the recent "Clog loss" Machine Learning competition that we recently co-ran with our partners, DrivenData. GAIA πŸ€– Β is joining Stall Catchers now as part of our ongoing research on hybrid AI/human systems.

The idea is to let GAIA πŸ€– analyze all the "easier" vessels, which she will do much faster than us humans, but with high enough accuracy that we need in our Stall Catchers data analysis. In the meantime, the "harder" movies we'll be left to us, for the irreplacable keen human eye to analyze! You can read more about the rationale of this approach and our hybrid AI research in this blog post.

We'll be posting more info about GAIA πŸ€– here on the blog soon!

Which dataset will we analyze?

For this catchathon, we are going to re-analyze the "Structural mapping" dataset, and see if we can do it just as well with our bot friend GAIA πŸ€– , as we did on our own!

This will help us:

  • validate our hybrid approach: it will show us whether the crowd answers we get with the bot are reliable enough, and let us know if we can continue to use GAIA πŸ€– Β on Stall Catchers for faster data analysis!
  • provide more confidence on the results already acquired for the "Structural mapping" dataset, as these results are not yet published and could use another "look" before the lab draws final conclusions
  • "Structural mapping" movies were taken with 3-photon microscopy - which is different to most other datasets in Stall Catchers (which use 2-photon microscopy) - and are of high quality, so it will be interesting to see how well the bot (which was trained on 2-photon microscopy data) will perform on these data

Can we reliably analyze a whole dataset in just one sitting?! Let’s see!

See you on Stall Catchers in a few hours! πŸ’œ