Official repo of the paper Deep Regression Unlearning accepted in ICML 2023.
We study the unexplored field of unlearning in deep regression models and introduce two unlearning methods, blindspot unlearning and gaussian amnesiac unlearning. We adapt metrics from earlier studies on unlearning in deep classification problems for the regression setting. We evaluate our method using a variety of unlearning metrics including different privacy attacks. We conduct regression experiments for computer vision, natural language processing and forecasting applications.
@article{tarun2022deep, title={Deep Regression Unlearning}, author={Tarun, Ayush K and Chundawat, Vikram S and Mandal, Murari and Kankanhalli, Mohan}, journal={arXiv preprint arXiv:2210.08196}, year={2022} }