Jeremy Here from Zama, if you are interested in learning more, check out our code on Github, everything we do is open source: https://github.com/zama-ai/fhevm
Zama's research on the application of Fully Homomorphic Encryption (FHE) to Large Language Models is now featured on @huggingface. Their Machine Learning team recently discovered that FHE has the potential to safeguard both user privacy and the Intellectual Property of the model.
Yes, you can find more information in this blog post, where we explain our choice to create a separate FHE library from our Concrete framework.
> https://www.zama.ai/post/announcing-tfhe-rs
Its main ambition is to show that FHE can be used for protecting data when using a Machine Learning model to predict outcomes without degrading its performance.
Disclaimer: I'm working at Zama (cited in the article posted).
While the idea behind Fully Homomorphic Encryption has been around for decades, it suffered three major issues: it was too slow, too hard to use, and too limited in terms of what you could do with it.
With Concrete, all this changes! Learn how we are addressing these three challenges in our official product release post.
And feel free to ask us if you have any question about FHE or what we do at Zama :)
Our main idea here was not only to build a predictive model that answers the question: “what sorts of people were more likely to survive?” but also to do it on encrypted data.
This was possible thanks to our Python package: Concrete-ML that aims to simplify the use of fully homomorphic encryption (FHE) for data scientists.
Feel free to ask the team if you have any questions about our work!
It's still early, but we do think that FHE is getting ready for prime time, and for being used in cloud applications. We are confident that it can be speed up by a factor of 100–1000x in the next 5 years, making FHE an ubiquitous technology for protecting privacy....
The next FHE.org meetup will be on November 23rd at 6pm Paris time. Rosario Cammarota from Intel will be talking about hardware acceleration for homomorphic encryption.
All the info are on the fhe.org website, including the link to register.
Finally, we have started a community discord to discuss about FHE, for those interested in continuing the conversation after the meetups! https://discord.gg/fvZ48443zD