Just a Simple Transformation is Enough for Data Protection in Vertical Federated Learning

Andrei Semenov, Philip Zmushko, Alexander Pichugin, Aleksandr Beznosikov
12/16/2024
cs.LGcs.CR

No official code found. Best implementation: Andron00e/JAST (1 star, Apache-2.0, last commit 2025-09-24).

Abstract

Vertical Federated Learning (VFL) aims to enable collaborative training of deep learning models while maintaining privacy protection. However, the VFL procedure still has components that are vulnerable to attacks by malicious parties. In our work, we consider feature reconstruction attacks, a common risk targeting input data compromise. We theoretically claim that feature reconstruction attacks cannot succeed without knowledge of the prior distribution on data. Consequently, we demonstrate that even simple model architecture transformations can significantly impact the protection of input data during VFL. Confirming these findings with experimental results, we show that MLP-based models are resistant to state-of-the-art feature reconstruction attacks.

View on arXivView PDF

Code Implementations(1)

Andron00e/JASTCommunity100%
Apache-2.0

''Just a Simple Transformation is Enough for Data Protection in Vertical Federated Learning''

13Python, Jupyter NotebookAug 19, 20241 years agoApache-2.0
feature-reconstructionprivacysplit-learningvertical-federated-learning

Similar Papers

UIFV: Data Reconstruction Attack in Vertical Federated Learning

92% match2024

BadVFL: Backdoor Attacks in Vertical Federated Learning

88% match1 repo232023

Practical and General Backdoor Attacks against Vertical Federated Learning

88% match2023

Mutual Information Regularization for Vertical Federated Learning

87% match2023

Privacy Against Agnostic Inference Attacks in Vertical Federated Learning

87% match2023

Defending Against Data Reconstruction Attacks in Federated Learning: An Information Theory Approach

87% match2024

Cite this paper

@article{semenov2024just,
  title  = {Just a Simple Transformation is Enough for Data Protection in Vertical Federated Learning},
  author = {Andrei Semenov and Philip Zmushko and Alexander Pichugin and Aleksandr Beznosikov},
  year   = {2024},
  eprint = {2412.11689},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2412.11689v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2412.11689.svg)](https://codeofpaper.com/paper/2412.11689)

Discussion