Skip to content

[ACM Computing Survey 2025] Vertical Federated Learning for Effectiveness, Security, Applicability: A Survey, by MARS Group at Wuhan University.

Notifications You must be signed in to change notification settings

shentt67/VFL_Survey

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

5 Commits
 
 
 
 
 
 

Repository files navigation

Vertical Federated Learning for Effectiveness, Security, Applicability: A Survey

Mang Ye, Wei Shen, Bo Du, Eduard Snezhko, Vassili Kovalev, Pong C. Yuen

Wuhan University, National Academy of Sciences of Belarus, Hong Kong Baptist University

Abstract Vertical Federated Learning (VFL) is a privacy-preserving distributed learning paradigm where different parties collaboratively learn models using partitioned features of shared samples, without leaking private data. Recent research has shown promising results addressing various challenges in VFL, highlighting its potential for practical applications in cross-domain collaboration. However, the corresponding research is scattered and lacks organization. To advance VFL research, this survey offers a systematic overview of recent developments. First, we provide a history and background introduction, along with a summary of the general training protocol of VFL. We then revisit the taxonomy in recent reviews and analyze limitations in-depth. For a comprehensive and structured discussion, we synthesize recent research from three fundamental perspectives: effectiveness, security and applicability. Finally, we discuss several critical future research directions in VFL, which will facilitate the developments in this field.

Survey for Vertical Federated Learning, by MARS Group at Wuhan University, led by Prof. Mang Ye.

contributions welcome

Table of Contents

Our Works

Survey

Generalized Federated Learning

Robust Federated Learning

Personalized Federated Learning

Federated Graph Learning

Vertical Federated Learning Survey

Survey Outline

Overview

Preliminary

Example of VFL

example_VFL

We present a practical cross-domain collaboration with three participants: mall, video platform, and bank. The mall acts as the active client, collaborating with the video platform and the bank as passive clients. Each client holds the local features and models of the same users. The active client holds the task labels, e.g., whether buying the cigar. A global model is introduced to make the final prediction of the shared/aligned users by aggregating feature embeddings from all clients. With prediction results and labels, the gradients can be calculated for both global and local model updation. Besides, a third-party coordinator can be employed for secure communication and sample alignment.

The Training and Testing Flow

Train_Test

(a) During training, aligned sample embeddings are sent to the active client, where gradients are calculated based on task labels. The overall objective is to optimize for collaborative prediction. These gradients are then sent back to each client for model updating. (b) During testing, predictions on aligned samples are made utilizing the trained global and local models.

Basic Research Directions

☀️Effectiveness

Model Design

Model_Design

Tree-based Model

Neural Network-based Model

Feature & Client Selection

Feature_Client_Selection

Feature Selection

Client Selection

☀️Security

Privacy Leakage

Secure Alignment

Secure Embedding Transportation

Secure Gradient Transportation

Inference Attack

Inference_Attack

Feature Inference Attack

Label Inference Attack

Destructive Attack

Destructive_Attack

Backdoor Attack

Poison Attack

Defense

Defense Against Feature Inference Attack

Defense Against Label Inference Attack

Defense Against Destructive Attack

☀️Applicability

Limited Data

Limited_Data

Limited Aligned Samples

Limited Labels

Large Communication Burden

Communication_Asynchrony

Client Asynchrony

Future Directions

Effectiveness/Applicability and Security Trade-off

Effectiveness Facilitates Security and Applicability

No recent works and will be a critical direction in the future.

Open Issues

Practical Datasets
Robustness and Generalization

Generalization to Unfair Prediction Bias

VFL on Different Data Variants

Multi-Modal Data

Graph Data

VFL with Foundation Models

Please kindly cite the paper if it helps your research, thanks!

@article{ye2024vertical,
  title={Vertical Federated Learning for Effectiveness, Security, Applicability: A Survey},
  author={Ye, Mang and Shen, Wei and Snezhko, Eduard and Kovalev, Vassili and Yuen, Pong C and Du, Bo},
  journal={arXiv preprint arXiv:2405.17495},
  year={2024}
}

About

[ACM Computing Survey 2025] Vertical Federated Learning for Effectiveness, Security, Applicability: A Survey, by MARS Group at Wuhan University.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published