OpenMined & Safe Artificial Intelligence

OpenMined evening: 3 members of the OpenMined team are joining us!
This is not an event to miss! Speak to the OpenMined team, learn what this is all about, how you (yes, you!) can become a contributor to this initiative!!

Btw we recently partnered with Meetup Mates: https://meetup-mates.com. If you would like to attend this meetup but feel overwhelmed by the idea of going and networking by yourself, or you would just like some company, Meetup Mates is a great way to connect with like-minded people to go to meetups with!

Agenda:
18:00 - 18:30 PM - Food & drinks

18:30 - 19:10 PM - Andrew Trask - Building Safe Artificial Intelligence with OpenMined

19:10 - 19:50 PM - Alan Aboudib (NLP Team Lead at OpenMined) - NLP SyferText

19:50 - 20:30 PM - Adam J Hall - PyGrid Tutorial

More Talk Details:

1. Andrew Trask - Building Safe Artificial Intelligence with OpenMined

In this talk, you will learn about some of the most important new techniques in secure, privacy-preserving, and multi-owner governed Artificial Intelligence. The first section of the talk will present a sober, up-to-date view of the current state of AI safety, user privacy, and AI governance. Andrew will then continue to introduce several fundamental tools of technical AI safety: Homomorphic Encryption, Secure Multi-Party Computation, Federated Learning, and Differential Privacy. The talk will finish with an exciting demo from the OpenMined open-source project showing how to train a deep neural network while both the training data AND model are in a safe, encrypted state during the entire process.

Andrew Trask is a PhD student at the University of Oxford where he researches new techniques for technical AI safety. With a passion for making complex ideas easy to learn, he is also the author of the book Grokking Deep Learning, an instructor in Udacity's Deep Learning Nanodegree, and he authors a popular Deep Learning blog iamtrask.github.io. He is also the leader of the OpenMined open-source community, a group of over 3000 researchers, practitioners, and enthusiasts which extends major Deep Learning frameworks with open-source tools for technical AI safety (openmined.org).

2. Alan Aboudib (NLP Team Lead at OpenMined) - NLP SyferText

In this demo, Alan will introduce you to SyferText, OpenMined's new privacy-preserving NLP library written in Python and built on top of PySyft. He will walk you through a concrete code example that demonstrates how SyferText can be used to create a deep learning NLP model and to perform encrypted federated training on multiple private datasets residing on different machines, abstracted as one bigger dataset, without violating any privacy constraint imposed on those datasets.
You will start by learning how to use the library to perform blind preprocessing on text you are not allowed to see due to such privacy concerns. Then, you will learn to build a deep learning model with PySyft that can be trained on an encrypted version of this data. Finally, you will see how you can use SyferText to encapsulate this model and add it to a pipeline of other encrypted models to use it to perform predictions on new private text data.

Alan is leading the NLP team at OpenMined, focusing on building SyferText, a privacy-preserving NLP library. He obtained a PhD in computer vision from Télécom Bretagne in France and he is currently the Head of Computer Vision at The Contillery; a Parisian startup aiming at providing quantitative metrics for evaluating the engaging power of visual social advertisement content. He is passionate about studying the role of visual attention in video action recognition, a project he had started as a post-doctoral researcher at Collège de France in 2018.

3. Adam J Hall - PyGrid Tutorial
   Details to follow