Antchain-MPC is an open-source library for providing multiparty computation (MPC) algorithms/protocols. It includes sec_softmoid (ACSAC'23), ...
sec_softmoid: Secure Softmax/Sigmoid for Machine-Learning Computation (ACSAC'23). [Paper] [Code] [Conference Slides]
Others: Coming later.
Sep 29, 2023
"Secure Softmax/Sigmoid for Machine-Learning Computation" has been accepted to ACSAC'23.
Sep 3, 2022
MPC-Secure KMeans, Partially-Secure PCA, Fully-Secure PCA have been updated. Running examples can be seen here. Please note that fake inverse_sqrt (used in Fully-Secure PCA) is NOT secure. We will release the real code later due to business reasons.
Morse-STF is a tool for privacy-preserving machine learning using MPC. We outline more information below:
| Folder | Description |
|---|---|
| conf | configuration files |
| ducuments | descriptive files (keep updating) |
| examples | running examples for stensorflow |
| stensorflow | importing folder for all functionality |
| unitest | testing files for different units used in stensorflow |
| output | store output files |
If you do not prefer to git clone our repository, python packages are here via pip install.
Part of code has been used in AntChain products. If you use it for business, please don't hesitate to get in touch with our team.
For academic usage, please cite our paper (under anonymous review, to be updated).