Machine Unlearning via Information-Theoretic Regularization
Published in Manuscript, 2025
Summary
Machine unlearning asks: can we remove specific training information from a trained model without retraining from scratch?
This manuscript proposes an information-theoretic approach to unlearning objectives and regularizers, emphasizing measurable behavior and reliable evaluation.
Resources
- Preprint (PDF): https://www.arxiv.org/pdf/2502.05684
- Code: (add link when public)
Recommended citation: Shizhou Xu, Thomas Strohmer. (2025). “Machine Unlearning via Information-Theoretic Regularization.” under review at Mathematical Foundations of Machine Learning.
Download Paper | Download Slides