The current vertical federated logistic regression uses homomorphic encryption (HE) to handle the privacy issues when multiple data owners share the data to train one model. However, homomorphic encryption cannot fully solve the privacy problem, because it requires a trusted server to aggregate all the data owners’ gradients to update a global model.
During the process, the trusted server must see all the data owners’ gradients in plaintext and send the plaintext global gradient back to each data owner. If a trusted server is attacked or dishonest, the data owners have a risk of leaking their data.
This invention uses MPC-SS to learn the model without a trusted server as well as keep the data privacy.
SoterOne designed a solution for multi-party collaboratively learning a logistic regression under Multi-Party Computation, Secret Sharing scheme, and it secures against an honest-but-curious adversary. The implementation can achieve the same accuracy as a naive non-private learning on logistic regression when all data are in one place, and it can scale to industry level data size.
Three main concepts about the invention: Logistic regression, Vertical federated learning and Secret sharing
Logistic regression is a well-known supervised machine learning method, and it is used to predict the probability of two or more classes such as win/lose, like/dislike. Logistic regression is easy to be trained, explained and has a solid statistical support. It is widely used in various fields, including machine learning, most medical, engineering and social sciences.
Vertical federated learning is two or more data owners would like to jointly learn a model by sharing their data under a privacy preserved environment. In this setting, each data owner maintains their own private data of different features about common entities(users, items), so that each data owner contributes to train a global model from locally computation on local data. In the vertical setting, the data is split by features(vertically), only one data owner has the target variable, and each data owner does not know the common entities across all the data owners.
Secret sharing: Secret sharing is a method to split a secret into multiple partitions and allocate a share of the secret to one party. The secret only can be reconstructed when a sufficient number of parties combine their share together.
Contributions of privacy preserved federated logistic regression
- We remove the dependence of the trust server, it allows joint learning to happen among multiple data owners without exposing data in plaintext, only the data owners itself can access its own data.
- The aggregate computation is on the plaintext powered by secret sharing, which is more efficient than computing on the ciphertext.
On the road to protect privacy, SoterOne will continue to explore more possibilities.
You can find us from the following channels:
Official website: www.soterone.com
Twitter: https://twitter.com/SoterOneBlock
Discord: https://discord.com/channels/811948111118598165/811948111979216938
