$400k for Data Insights; ML Privacy; JetBrains Giveaway

ML Contests: October 2020 Update


Hello again - or if you're one of our many new subscribers, welcome!

I hope you're doing well and that you've managed to find some good ML competitions to get stuck into if, like many of us, you're spending more time indoors than usual.

Since the last email went out, there are many new competitions - with total additional prize money of more than $600k! As always I'll highlight some of the most exciting competitions here, and you can go to mlcontests.com for the full list. Let's get stuck in...


Data Insights


Many recent competitions are a little different to the standard Kaggle supervised learning problem. Rather than getting a dataset to label and being scored programmatically, these competitions are open-ended, asking you to look over large amounts of (often unstructured) data, and come up with meaningful insights. As well as data skills, you need to be able to come up with interesting ideas and communicate your solution well. There are three notable competitions in this vein:

- COVID-19 Grand Challenge ($200k in prizes + accelerated recruiting process at C3.ai): C3.ai challenge you to come up with "meaningful data-driven insights that will fuel leading decisions to fight the COVID-19 pandemic". You get REST API access to their huge COVID-19 Data Lake which comprises daily case reports, epidemiology data, genome sequences, journal articles, clinical patient data, policy actions, mobility statistics, demographics, economic indicators, environmental factors, and public surveys from around the world. There are so many directions you could go with this, and as well as very generous prizes there's the potential to make a difference to the course of this pandemic (in contrast, the FT recently wrote how "AI has disappointed on COVID"). Registration closes on 25th Oct, and the submissions deadline is 18th Nov.

- Unlocking Climate Solutions ($91k in prizes): CDP challenge you to use their world-class environmental action action dataset to come up with key performance indicators for important environmental and social issues like climate change, water security, deforestation, and social inequity. Final submissions deadline is 2nd Dec 2020.

- NFL Big Data Bowl 2021 ($100k in prizes): a more light-hearted competition. Use NFL's crazy detailed Next Gen Stats data to come up with a way to measure defensive performance across drop-back pass plays from the 2018 regular season. The data you get access to includes the position and speed of every player on the field during each play, so anything is possible! Final submissions deadline is 7th Jan 2021.


Differential Privacy


Another exciting new competition run by NIST and NASA, the Differential Privacy Temporal Map Challenge ($147k in prizes) asks: how can we best process location data in a way that protects the privacy of those present in the data, but maintains the data's usefulness in applications like natural disaster response and epidemic tracking? There's a clear trade-off here between privacy and utility, and your challenge is to develop algorithms that preserve data utility as much as possible while guaranteeing individual privacy is protected. Time to dust off that information theory book! There are three 'sprints' in this challenge, running Oct-Dec, Jan-Mar, and Apr-Jun.

Thank You - JetBrains Giveaway


ML Contests is a community project, and a special thank you goes out to everyone who's contributed on GitHub to keep the website updated: Himanshu Garg, Anton Pankratov, Shivam Khandelwal, Vrushank Vyas, Anish Jain, Naman Bhardwaj, and Ram Dewani!

As a way of saying thanks to all of you who've subscribed, we've persuaded JetBrains to gift year-long licences for their software to two readers of this newsletter. JetBrains are the makers of PyCharm - which, with its Scientific Mode and other useful features, is the tool of choice for many Python ML researchers. Their other tools include Clion for C++ and IntelliJ IDEA for Java.



In order to be in with a chance of winning one of these licences worth $649, follow @ml_contests on Twitter and mention @ml_contests in a tweet before midnight GMT this Sunday, 25th October 2020. Winners will be chosen randomly, and notified via DM.

Good luck! And if you know anyone who might enjoy this newsletter, please forward it on :)

Harald @ ML Contests
mlcontests.com | harald.co