Sofiane OUAARI

Join us for this conference to explore the importance, as well as the "what," "how," and "why" of privacy-preserving machine learning (ppML). At the intersection of data regulations and secure machine learning systems, ppML enables us to enhance the privacy of our machine learning models against various types of privacy attacks, such as membership inference, model inversion, and data poisoning. This talk will cover multiple ppML solutions, including differential privacy, controlled representation learning, and federated learning.