Original title: SecureCut: Federated Gradient Boosting Decision Trees with Efficient Machine Unlearning
Authors: Jian Zhang, Bowen Li Jie Li, Chentao Wu
In the world of data privacy, a new mandate demands the erasure of user data, prompting a challenge: enabling data removal in Vertical Federated Learning (VFL). Enter a novel solution named \methname. This innovative Gradient Boosting Decision Tree (GBDT) framework allows the removal of specific features, ensuring privacy across multiple parties contributing to model training.
Original article: https://arxiv.org/abs/2311.13174