Encrypted Machine Learning Inference

"This study investigates the use of Fully Homomorphic Encryption (FHE), a technique that allows operations to be performed over encrypted data, with machine learning models. We evaluate the feasibility of deploying these models in realistic computational constraints and measure the accuracy of inference over plain text compared to encrypted data. For that purpose, we utilize Zama Concrete-ML library and implement classification and regression models. Our goal is to assess the practicality of using FHE based machine learning models."

5:30-- Networking at STARTUP COLUMBUS
6:00-- Presentation and Discussion
7:17-- Discussion at NONIC in COLUMBUS, GA

• Important to know
We normally use the first 30 min to socialize a bit and will start the presentation promptly at 6pm. The group meets for drinks & discussion at Nonic after the meetup. All are welcome to attend.