Facebook offers 5,000 scholarships to study privacy-friendly artificial intelligence techniques.
As Facebook announced this week, within the framework of its F8 developer conference, the social network company has joined forces with the e-learning portal Udacity to promote privacy-friendly artificial intelligence techniques.
And it has done so by offering 5,000 scholarships for people interested in take the ‘Secure and Private AI’ course for 3 months (more information here). And, once the course is finished, the best 300 students will have the opportunity to take a ‘nanodegree’ (on ‘Deep Learning’ or on ‘Computer Vision’).
Why is Facebook interested in promoting privacy-friendly use of AI?
Facebook is, today, a company that makes intensive use of deep learning and that has on its staff pioneers in the field of artificial intelligence like the recently awarded Yann LeCun. And yet the majority of the public increasingly tends to link their brand to privacy concerns.
Scandals like Cambridge Analytica still weigh, and Zuckerberg himself acknowledged this week that he knows that “Right now Facebook doesn’t exactly have the strongest reputation for privacy“Yet he insisted on his new mantra:” The future will be private. “
AND Facebook wants to be part of it. That is precisely where this new course comes in: a sequel to a previous one that dealt with the deep learning framework developed by Facebook AI, PyTorch, and that already attracted over 18,000 students at the end of last year.
Andrew Trask, researcher at DeepMind, and leader of the OpenMined project (to develop private machine learning software) explains the importance of a course like this, which he himself has helped to design:
“Without data scientists who know how to properly handle privacy, private data will in many cases go unused (a critical loss in fields like healthcare) or exposed by resorting to techniques that lack adequate privacy protections.”
Certainly, some of the most important uses that we can give AI today require sufficient knowledge about the use of private data, a field in which they have been given. important steps in recent years.
Therefore, the Udacity and Facebook course provides specific training on three key technologies: differential privacy, federated learning and encrypted computing. The software tools to be used during the course will be the aforementioned PyTorch and its complement PySyft, developed by the OpenMined project.
How to get the scholarship
As Udacity explains, the scholarship is open to all candidates over 18 years of age who have some experience using Python and deep learning.
The candidates they have until May 21 to apply for the scholarship (The process is carried out on the same website as the course), and if they are selected they will be notified the same day the course opens, next May 30.