Homomorphically Encrypted Deep Learning Library
The goal of this library is to give the user the ability to efficiently train machine learning models in a homomorphically encrypted state without needing to be an expert in either. Furthermore, by understanding the characteristics of both Deep Learning and Homomorphic Encryption, we hope to find very performant combinations of the two. See notebooks folder for tutorials on how to use the library.
You need to install this library locally before running any of the notebooks this repository or the main demonstration:
# Get dependencies ready
pip install -r requirements.txt
# install the lib locally
python setup.py install
bash install_for_anaconda_users.sh
Install Docker from https://www.docker.com/
For macOS users with Homebrew installed, use brew cask install docker
Then, run:
git clone https://github.com/OpenMined/PySyft.git
cd PySyft/notebooks/
docker run --rm -it -v $PWD:/notebooks -w /notebooks -p 8888:8888 openmined/pysyft jupyter notebook --ip=0.0.0.0 --allow-root
If you are interested in contributing to Syft, first check out our Contributor Quickstart Guide and then checkout our Project Roadmap and sign into our Slack Team channel #syft to let us know which projects sound interesting to you! (or propose your own!).
As both Homomorphic Encryption and Deep Learning are still somewhat sparsely known, below is a curated list of relevant reading materials to bring you up to speed with the major concepts and themes of these exciting fields.
- How to build a fully encrypted AI model (trained on unencrypted data):
- Simple secure protocol for federated machine learning (using a python-paillier library):
- Prototype for using encrypted AI to preserve user privacy (in python):
- Manual for Using Homomorphic Encryption for Bioinformatics (paper):
- https://eprint.iacr.org/2014/062.pdf
- http://heat-h2020-project.blogspot.co.uk/2017/02/homomorphic-encryption-api-software.html
- http://heat-h2020-project.blogspot.co.uk/2017/
- http://proceedings.mlr.press/v54/esperanca17a/esperanca17a.pdf
- https://sites.google.com/view/psml/program/abstracts
- https://arxiv.org/pdf/1508.06845.pdf
- https://arxiv.org/pdf/1508.06574.pdf
- https://eprint.iacr.org/2016/892
- https://www1.cs.fau.de/filepool/publications/octavian_securescm/secfp-fc10.pdf
- http://www.louisaslett.com/HomomorphicEncryption/#details
- https://github.com/iamtrask/linreg-mpc
- https://github.com/dask/dask-tutorial
- https://mortendahl.github.io/2017/04/17/private-deep-learning-with-mpc/
- https://mortendahl.github.io/2017/06/04/secret-sharing-part1/
- https://mortendahl.github.io/2017/06/24/secret-sharing-part2/
- https://blog.n1analytics.com/distributed-machine-learning-and-partially-homomorphic-encryption-1/
- https://blog.mturk.com/tutorial-how-to-verify-crowdsourced-training-data-using-a-known-answer-review-policy-85596fb55ed