Exact Machine Unlearning
摘要
From the previous chapter, we have learned that machine unlearning (MU) is a critical mechanism to uphold users’ right to be forgotten. One representative unlearning is exact machine unlearning, which is extended from naive retraining from scratch method. Exact unlearning methods design training procedures and model structures that allow specified samples to be removed without fully repeating the retraining process. Through the study of this chapter, readers will have a basic understanding of exact machine unlearning.