Privacy Preservation through Practical Machine Unlearning
This work addresses privacy preservation for AI systems, but it is incremental as it builds on existing unlearning methods and frameworks.
The paper tackles the problem of selectively removing data from trained machine learning models to address privacy concerns, evaluating methods like Naive Retraining and SISA on the HSpam14 dataset and finding that unlearning frameworks like DaRE can ensure privacy compliance with significant computational trade-offs.
Machine Learning models thrive on vast datasets, continuously adapting to provide accurate predictions and recommendations. However, in an era dominated by privacy concerns, Machine Unlearning emerges as a transformative approach, enabling the selective removal of data from trained models. This paper examines methods such as Naive Retraining and Exact Unlearning via the SISA framework, evaluating their Computational Costs, Consistency, and feasibility using the $\texttt{HSpam14}$ dataset. We explore the potential of integrating unlearning principles into Positive Unlabeled (PU) Learning to address challenges posed by partially labeled datasets. Our findings highlight the promise of unlearning frameworks like $\textit{DaRE}$ for ensuring privacy compliance while maintaining model performance, albeit with significant computational trade-offs. This study underscores the importance of Machine Unlearning in achieving ethical AI and fostering trust in data-driven systems.