APNNS/IEEE Education Forum Series: Deep Learning and Artificial Intelligence Summer School 2020
(DLAI3)(https://deeplearningandaiwinterschool.github.io/)が29 June – 3 July 2020, Bangkok, Thailand +
Academic Talk 1 01.15 – 02.15 pm.
Speaker: Seiichi Ozawa, Kobe University, Japan
Topic: An Introduction to Privacy-Preserving Machine Learning for Big Data Analysis
Since the advancement of AI brings us various smart services that are strongly linked to our lives, the value of personal data is rapidly increasing year by year. On the other hand, there is also an increasing risk that privacy invasion and leakage could cause serious damages to ordinal users. Concerning such privacy risk, even though a secure cloud computing environment is available for analyzing such sensitive data, it might not be accepted by data holders because data could be viewed by a cloud owner, so-called “semi-honest setting”. Therefore, the expectation for new data processing technologies that can analyze data securely are raising recently. This tutorial presents some of the latest technologies for privacy-preserving data analysis that enables us to analyze data securely while protecting privacy. In particular, I will give a brief explanation on k-anonymity, differential privacy, privacy-preserving machine learning using homomorphic encryption, and federated learning that allows us to analyze/share personal information on edge devices without revealing the contents each other.