/TrustworthyFederatedLearning

THU-AIR 联邦学习隐私与安全

Primary LanguagePython

Trustworthy Federated Learning

This repository is the collection of code for all related works on Trustworthy Federated Learing by FL-AIR group. Each folder under this repository is a set of experiments that focus on several closely related attacks. And each target to a paper published. Detail information can be found in those README.md files under each folder.

1. Folders List

Folder Setting Attack Defense Paper
LabelDefender VFL(2-party or 4-party) Batch-level Label Inference Attack, Label Replacement Backdoor Attack CAE, DCAE Defending Batch-Level Label Inference and Replacement Attacks in Vertical Federated Learning
PairedLogitsInversion FL (multi-party) Paired Logits Inversion (PLI) - Breaching FedMD: Image Recovery via Paired-Logits Inversion Attack

2. Future Updates

This folder will be updated in the future by adding more relavent code for published paper focusing on VFL privacy and safety.