For our third Zero Knowledge event, on January 23th, we are pleased to host two great speakers: Vladislav Gelfer and a special guest coming directly from NeurIPS, Morten Dahl.
We will be talking about Cryptographic Private Machine Learning and Privacy on Blockchain.
Schedule:
- 6.30pm: greetings, food and drinks
- 7pm: Vladislav Gelfer will present MimbleWimble high level concepts and existing implementations.
- 7.30pm: Morten Dahl will give a talk on Privacy Technologies for Machine Learning
The event will finish at 8.30pm.
Talks will be given in English.
For more details, here the abstracts and bios of our speakers:
In this talk we focus on recent applications of advanced cryptographic methods to machine learning, in particular deep learning. After illustrating how tools such as homomorphic encryption and multi-party computation can benefit the machine learning process in terms of privacy and trust, we proceed to give a high-level overview of their underlying principles in order to understand differences, weaknesses, and strengths. As an example we show how a model can be trained on data that remain encrypted throughout the whole process. We do so using tf-encrypted, a library on top of TensorFlow for working with encrypted data.
Morten holds a PhD in cryptography and works in the intersection of privacy and machine learning. He is interested in practical tools and concrete applications, with a current focus on making advanced privacy-enhancing tools more accessible to practitioners. He is a recurrent speaker in the field and active in community building.
