Control, Confidentiality, and the Right to be Forgotten in Differential Privacy for Machine Learning
Explore the complex intersection of data privacy, machine learning, and user rights in this Google TechTalk presented by Aloni Cohen. Delve into the challenges of data deletion, confidentiality, and the right to be forgotten in the context of machine learning systems. Examine the concept of differential privacy and its application to ML, focusing on the problems associated with retraining models after data deletion. Investigate Twitter's approach to data deletion and the implications for user privacy. Analyze the relationship between control and confidentiality in data management, and explore the concepts of simulatable deletion and history independence. Gain insights into the theoretical foundations of data privacy, including randomness, coupling, and adaptive history independence. Conclude with a comparison of differential privacy and machine learning approaches to data protection.