Scissorhands: Scrub Data Influence via Connection Sensitivity in Networks

arXiv:2401.06187v2 Announce Type: replace
Abstract: Machine unlearning has become a pivotal task to erase the influence of data from a trained model. It adheres to recent data regulation standards and enhances the privacy and security of machine learning applications. In this work, we present a new machine unlearning approach Scissorhands. Initially, Scissorhands identifies the most pertinent parameters in the given model relative to the forgetting data via connection sensitivity. By reinitializing the most influential top-k percent of these parameters, a …

abstract applications arxiv become cs.cv cs.lg data data regulation influence machine machine learning machine learning applications networks pivotal privacy privacy and security regulation security sensitivity standards type unlearning via work