Machine unlearning via GAN

Abstract

Machine learning models, especially deep models, may unintentionally remember information about their training data. Malicious attackers can thus pilfer some property about training data by attacking the model via membership inference attack or model inversion attack. Some regulations, such as the EU's GDPR, have enacted "The Right to Be Forgotten" to protect users' data privacy, enhancing individuals' sovereignty over their data. Therefore, removing training data information from a trained model has become a critical issue. In this paper, we present a GAN-based algorithm to delete data in deep models, which significantly improves deleting speed compared to retraining from scratch, especially in complicated scenarios. We have experimented on five commonly used datasets, and the experimental results show the efficiency of our method.

Sentiment analysis and opinion mining on educational data: A survey

“…• Model optimization for the computational resources used by unlearning techniques by using techniques such as model compression, model quantization or model approximation. These techniques can reduce the computational overhead of unlearning methods and enable their deployment in resource-constrained environments or on lowpower devices [19], [199].…”

Fast and Accurate SNN Model Strengthening for Industrial Applications

“…A simple and naive solution is to delete the malicious data samples from the training data and retrain the model, which is usually quite time-consuming due to the huge computational load. There are several solutions to address this problem in an efficient way [14][15][16][17]. For example, Cao [18] divided the training data into several groups, removed the group with malicious data samples, and updated the whole model accordingly.…”