Multi-VFL: A Vertical Federated Learning System for Multiple Data and Label Owners

TL;DR

This paper introduces Multi-VFL, a novel vertical federated learning system accommodating multiple data and label owners, utilizing split learning and adaptive optimizers to enhance convergence and accuracy without data sharing.

Contribution

It is the first to address multi-owner data and label distribution in vertical federated learning, proposing a new framework with split learning and adaptive optimizers.

Findings

Adaptive optimizers improve convergence speed.

Model accuracy is enhanced with adaptive optimizers.

Experiments on MNIST and FashionMNIST validate effectiveness.

Abstract

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.