Brooke Joseph
Privacy in AI: A look into Federated Learning on TensorFlow
Saturday, 20 May 2023 11:05 am - 11:20 am (15 minutes)
Virtual Talk
Festival Room
Embedded Files
We will have a look into Federated Learning and how to apply Differential Privacy to further secure the process and the data. Walk through a real example of how to use TensorFlow Federated to make a real Federated model using the enmist dataset and apply Differential Privacy with TensorFlow Privacy.
Brooke is specifically interested in looking at the accuracy/privacy tradeoff that comes with DP, and how we can work to maximize both privacy and model utility in TensorFlow Privacy and how Federated Learning addresses the challenges that come with GDPR.
About the Speaker
The Knowledge Society, Innovator
Brooke is an 18 year old Innovator from Canada. She is extremely passionate about AI and learning about privacy-preserving technology. Her main interest is in Federated Learning, which is a new AI setting introduced by Google in 2017. She was able to develop a few projects in the space with the help from mentors at The Knowledge Society. This is a human competitive accelerator program that accepts the brightest teens in tech to work on solving some big problems. From this experience, she also had the opportunity to work alongside Walmart and Amazon to pitch AI solutions for integrating technology into their businesses. She has also been able to research at TensoAI and has had her work published on BetterProgrammer's page.
Page updated
Google Sites
Report abuse