Ant Group, the finance arm of Alibaba, has made its privacy-preserving technology SecretFlow open source with a focus on data analysis and machine learning. A secure processing unit, which offers secure computation capabilities guaranteeing data privacy, a homomorphic encryption unit, a portable simplest oblivious transfer protocol implementation, and SecretFlow, a higher-level unified framework integrating all of them, are some of the components that make up SecretFlow. Lower-level modules are written in C, C++, and assembly while the higher-level SecretFlow module is written in Python.
Completeness, transparency, openness, and interoperability with other technologies are goals of SecretFlow. The framework, according to the Ant Group, intends to facilitate the development of privacy-preserving computing-based apps by developers and to contribute to the continued development of the market and technological maturity. Pip install -U secretflow can be used to install SecretFlow. The snippet that follows demonstrates how to produce a random number between 3 and 4 for a particular user in standalone mode:
import secretflow as sf
>>> sf.init([‘alice’, ‘bob’, ‘carol’], num_cpus=8, log_to_driver=True)
>>> dev = sf.PYU(‘alice’)
>>> import numpy as np
>>> data = dev(np.random.rand)(3, 4)
SecretFlow can also be set up in cluster mode, increasing privacy by dedicating nodes to individual users. Ray, an open source framework that offers a straightforward, all-purpose API for developing distributed applications, serves as the foundation for SecretFlow cluster mode. Check out the tutorials for a short introduction to SecretFlow.
They cover a variety of use scenarios, such as data preprocessing, logistic regression, neural network training, and more. A method called privacy-preserving computation seeks to protect private information during processing. You can do computation on encrypted data using such methods, such as homomorphic encryption, which ensures that it cannot be captured or altered while being processed.