Multi-VFL: A Vertical Federated Learning System for Multiple Data and Label Owners
This addresses the challenge for entities with split data and labels to collaborate on model training without sharing data, though it is incremental as it extends existing VFL concepts.
The paper tackles the problem of training vertical federated learning models when both features and labels are distributed across multiple owners, proposing Multi-VFL as a novel method. Results on MNIST and FashionMNIST datasets show that using adaptive optimizers speeds up convergence and improves accuracy.
Vertical Federated Learning (VFL) refers to the collaborative training of a model on a dataset where the features of the dataset are split among multiple data owners, while label information is owned by a single data owner. In this paper, we propose a novel method, Multi Vertical Federated Learning (Multi-VFL), to train VFL models when there are multiple data and label owners. Our approach is the first to consider the setting where $D$-data owners (across which features are distributed) and $K$-label owners (across which labels are distributed) exist. This proposed configuration allows different entities to train and learn optimal models without having to share their data. Our framework makes use of split learning and adaptive federated optimizers to solve this problem. For empirical evaluation, we run experiments on the MNIST and FashionMNIST datasets. Our results show that using adaptive optimizers for model aggregation fastens convergence and improves accuracy.