The paper establishes that the optimal excess risk for ε-unlearning is the usual statistical error plus an unlearning penalty that interpolates between retraining-from-scratch and an exponentially smaller term as ε/d grows, with matching bounds for mean estimation.
Proceedings of the 23rd ACM Conference on Economics and Computation , location =
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
verdicts
UNVERDICTED 2representative citing papers
PATCH model simulations show preferential attachment and homophily increase segregation and degree inequality while triadic closure reduces segregation but amplifies overall inequality, and the model accounts for observed gender disparities in 50 years of physics and CS collaboration networks.
citing papers explorer
-
Near-Optimal Pure Machine Unlearning for Smooth Strongly Convex Losses
The paper establishes that the optimal excess risk for ε-unlearning is the usual statistical error plus an unlearning penalty that interpolates between retraining-from-scratch and an exponentially smaller term as ε/d grows, with matching bounds for mean estimation.
-
Network Inequality through Preferential Attachment, Triadic Closure, and Homophily
PATCH model simulations show preferential attachment and homophily increase segregation and degree inequality while triadic closure reduces segregation but amplifies overall inequality, and the model accounts for observed gender disparities in 50 years of physics and CS collaboration networks.