The paper introduces an auditable, information-theoretic 'marginal unlearning' definition and a rate-distortion style regularization method that can remove data points or features from models.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Machine Unlearning via Information Theoretic Regularization
The paper introduces an auditable, information-theoretic 'marginal unlearning' definition and a rate-distortion style regularization method that can remove data points or features from models.