Vertical Federated Learning in Practice: The Good, the Bad, and the Ugly
This survey addresses the limited real-world adoption of VFL, a privacy-preserving collaborative learning method, by highlighting research-practice gaps for cross-organizational applications.
The paper investigates the gap between existing Vertical Federated Learning (VFL) research and practical deployment by analyzing real-world data distributions, proposing a data-oriented taxonomy, and identifying scenarios with few viable solutions.
Vertical Federated Learning (VFL) is a privacy-preserving collaborative learning paradigm that enables multiple parties with distinct feature sets to jointly train machine learning models without sharing their raw data. Despite its potential to facilitate cross-organizational collaborations, the deployment of VFL systems in real-world applications remains limited. To investigate the gap between existing VFL research and practical deployment, this survey analyzes the real-world data distributions in potential VFL applications and identifies four key findings that highlight this gap. We propose a novel data-oriented taxonomy of VFL algorithms based on real VFL data distributions. Our comprehensive review of existing VFL algorithms reveals that some common practical VFL scenarios have few or no viable solutions. Based on these observations, we outline key research directions aimed at bridging the gap between current VFL research and real-world applications.