New Facebook Scholarship Challenge On Udacity
Written by Sue Gee   
Thursday, 02 May 2019

At its F8 developer event this week, Facebook and Udacity announced a new scholarship program. It is based around a new course that aims to teach AI engineering skills in privacy-preserving technologies.

 

facebooksapaichallenge

This new online learning scholarship program is a continuation of the PyTorch Scholarship Challenge, the previous partnership between Udacity and Facebook announced at the PyTorch Developer Conference last October. That program had more than 18,000 applicants, 10,000 challenge scholars and led to 300 students from 149 countries being enrolled in the Udacity Deep Learning Nanodegree program with a full scholarship from Facebook.

This time around the Secure and Private AI Scholarship Challenge is at advanced level. The new course on Secure and Private AI  that aspiring scholarship students need to complete  has been built in collaboration with Andrew Trask, Leader of OpenMined and Research Scientist at DeepMind, who is lead instructor in the course. 

 

As Andrew Trask explains in the video, while privacy is essential, it hampers research by limiting access to human data. For this reason the course focuses in methods for the principled protection of user privacy using the newest privacy-preserving technologies, such as OpenMined's PySyft. PySyft extends Deep Learning tools—such as PyTorch—with the cryptographic and distributed technologies necessary to safely and securely train AI models on distributed private data.

The course, which is free of charge, will open on May 30th. It is self-paced and lasts approximately 2 months. It comprises three lessons:

  1. Differential Privacy

    Learn the mathematical definition of privacy
    Train AI models in PyTorch to learn public information from within private datasets

  2. Federated Learning

    Train on data that is highly distributed across multiple organizations and data centers using PyTorch and PySyft
    Aggregate gradients using a "trusted aggregator" 

  3. Encrypted Computation
    Do arithmetic on encrypted numbers
    Use cryptography to share ownership over a number using Secret Sharing
    Leverage Additive Secret Sharing for encrypted Federated Learning 

Facebook’s support is making 5,000 seats available for a supported version of the Secure and Private AI course. Upon successfully completing the first phase of the program, 300 students will then go on to earn full Nanodegree program scholarships for either of Udacity's Deep Learning Nanodegree program or its  Computer Vision Nanodegree program

As with the PyTorch Scholarship Challenge, applying to join the program involves completing questionnaires. In addition to Background Information applicants have to establish their Prerequisite Knowledge and outline the Goals. The deadline for this is May 21. 

 blueudacity

 

Banner


Quantum Computers Really Are A One Trick Pony
17/03/2024

Google is offering $5 million if you can think up a use for a quantum computer. Wait, I thought quantum computers were the next big thing, a revolution! Surely we know what they can do?



iOS 17.4 Released With Support For App Stores In The EU
06/03/2024

I have written about Apple's approach to complying with regulation, characterizing it as malicious compliance. It also seems that Apple is a master of creating the unintended consequence and letting i [ ... ]


More News

raspberry pi books

 

Comments




or email your comment to: comments@i-programmer.info

Last Updated ( Friday, 03 May 2019 )