Call for participation: NeurIPS 2020 hide-and-seek privacy challenge
Part of the NeurIPS 2020 competition track, the hide-and-seek challenge explores the meaning and limitations of privacy, offering participants a chance to develop and test synthetic data techniques that could revolutionize how clinical data is used in research.
The challenge takes the form of a game of hide-and-seek within a comprehensive, high-resolution intensive care database provided by Amsterdam UMC.
In the synthetic data generation track, “hiders” are tasked with developing an algorithm that generates synthetic data on the basis of real clinical data.
In the patient re-identification track, “seekers” must develop an algorithm that performs membership inference (patient re-identification) on the synthetic data generation algorithms created by hiders.
Duration and prizes
Microsoft Research is providing 2 cash prizes of $5,000 for the winning teams.
Teams entering the competition in July/August will have the chance to win $250 in Microsoft Azure credits.
Participants may enter between July and October 2020.
To enter, visit https://www.vanderschaar-lab.com/privacy-challenge/
Organizers
- van der Schaar Lab (University of Cambridge)
- Microsoft Research Lab - Cambridge
Other papers on privacy-preserving machine learning and synthetic data
https://www.vanderschaar-lab.com/publications/synthetic-data
Researcher @ UCF CRCV | PhD in Electrical Engineering
4yI will be joining after listening to your amazing talk today at OXML20
VP & Chief Data Scientist, Biopharmaceuticals R&D at AstraZeneca
4ySounds like a great challenge - looking forward to seeing how the competition goes! :)