Contents

Fully homomorphic encryption (FHE)

全同态加密 Fully homomorphic encryption (FHE)

What’s is FHE

第一种形式:同时使用secret key加密和解密,public key用于第三方加密进行同态计算

https://cdn.jsdelivr.net/gh/JoshuaChou2018/oss@main/uPic/image-20220810161455665.7Sogoe.png

第二种形式:asymmetric FHE,只使用public key进行加密,secret key用于解密

https://cdn.jsdelivr.net/gh/JoshuaChou2018/oss@main/uPic/image-20220810161707363.4nIJAt.png

Efficient FHE is a giant leap towards httpz://

? what’s httpz?

History

https://cdn.jsdelivr.net/gh/JoshuaChou2018/oss@main/uPic/image-20220810162634078.kd7s8b.png

Major task: find a new scheme to combine both addition and multiplication together.

1. First generation

Plain text: single bit

https://cdn.jsdelivr.net/gh/JoshuaChou2018/oss@main/uPic/image-20220810162938730.6YwAEf.png

这个方法的主要问题是noise,随着计算的叠加,there is a notion of noise in ciphertexts,noise accumulates

  1. Defining noise budget thourgh defining parameters

  2. noise growth is exponential

  3. bootstrapping to rescure (decrease the noise)

Somewhat homomorphic encyption (SHE)

初代的FHE目标是每一步同态计算,都进行一次bootstrapping,来抑制噪音增加。但是bootstrapping的计算消耗太大,所以SHE的目标是控制在output时的噪音水平,保证是可以接受的。

https://cdn.jsdelivr.net/gh/JoshuaChou2018/oss@main/uPic/image-20220810163706513.adONzi.png

2. Second generation

targeting the exponential growing noise

https://cdn.jsdelivr.net/gh/JoshuaChou2018/oss@main/uPic/image-20220810164134885.dqkLd9.png

https://cdn.jsdelivr.net/gh/JoshuaChou2018/oss@main/uPic/image-20220810164255110.L8cELM.png

3. Third generation: GSW

只是概念上的简化,但是无法实用。

https://cdn.jsdelivr.net/gh/JoshuaChou2018/oss@main/uPic/image-20220810164410551.9fFNfF.png

4. 4th generation: Torus FHE, TFHE

https://cdn.jsdelivr.net/gh/JoshuaChou2018/oss@main/uPic/image-20220810164528414.a1nVpv.png

Open-source FHE librabies

FV-NFLlib

SEAL

HEAAN

TFHE

HElib

nuFHE

Palisade

Two encryption

https://cdn.jsdelivr.net/gh/JoshuaChou2018/oss@main/uPic/image-20220810172343243.YM9jwM.png

Plaintext encoding

https://cdn.jsdelivr.net/gh/JoshuaChou2018/oss@main/uPic/image-20220810172439231.q2rqqZ.png

但是在计算的时候,noise会在右侧累积。

https://cdn.jsdelivr.net/gh/JoshuaChou2018/oss@main/uPic/image-20220810172529388.zPkVYt.png

通过bootstrapping,可以减少右侧的noise

https://cdn.jsdelivr.net/gh/JoshuaChou2018/oss@main/uPic/image-20220810172605531.e9V2vW.png

LWE ciphertexts are homomorphic

https://cdn.jsdelivr.net/gh/JoshuaChou2018/oss@main/uPic/image-20220810172707491.kX7bKI.png

Ref

Complete talk: Introduction to Homomorphic Encryption (by Pascal Paillier) - YouTube