#Privacy Preserving Machine Learning

Homomorphic Encryption

解釋

同態加密。可以在加密的情況進行數學運算

同態加密的類別

1. 完全同態加密（Fully Homomorphic Encryption，FHE）
2. 部分同態加密（Partially Homomorphic Encryption，PHE）
3. 特定同態加密（Somewhat Homomorphic Encryption，SHE）
4. 加法同態、乘法同態、減法同態、除法同態

awesome

1. jonaschn/awesome-he

reference

1. Types of Homomorphic Encryption
2. 同態加密

Differential Privacy

解釋

在數據中引入一些隨機性，使得無法確定單個人的貢獻。

差分隱私的類別

Multi-Party Computation

awesome

1. rdragos/awesome-mpc

Federated Learning

awesome

Split Learning

awesome

1. splitlearning/awesome-split-learning

Attack

Conferences and Workshops

OCW

Keywords

1. Ring learning with errors（RLWE）
2. Secure Multiparty Computation
   • https://digitalprivacy.ieee.org/publications/topics/applications-of-multiparty-computation

Open Source

1. SPU: Secure Processing Unit
   • https://github.com/secretflow/spu
   • SecretFlow-SPU: A Performant and User-Friendly Framework for Privacy-Preserving Machine Learning
2. facebookresearch/CrypTen
   • https://github.com/facebookresearch/CrypTen
   • A framework for Privacy Preserving Machine Learning
3. secretflow/secretflow
   • https://github.com/secretflow/secretflow

網路資源

1. Privacy-Preserving-Machine-Learning-Resources
2. awesome-ppml
3. awesome-privacy-preserving-machine-learning

社團

1. Cypherpunks-Core