SecretFlow is a unified framework for privacy-preserving data analysis and machine learning.
Device abstraction, which abstracts privacy computing technologies such as Multi-Party Secure Computing (MPC), Homomorphic Encryption (HE), and Trusted Execution Environment (TEE) into ciphertext devices, and abstracts plaintext computing into plaintext devices.
Computational graphs based on abstracted devices, enabling data analysis and machine learning workflows to be represented as computational graphs.
Machine learning/data analysis capabilities based on computational graphs, supporting data horizontal/vertical/hybrid segmentation and other scenarios.
At present, privacy computing technology is growing in popularity. However, neither the technology nor the market has yet reached real maturity. In order to cope with the development uncertainty of privacy computing technology and applications, we propose a general privacy computing framework called “SecretFlow”. SecretFlow will adhere to the following principles, so that the framework has the maximum inclusive and extensible capabilities to cope with the development of future privacy computing technologies and applications.
Completeness: It supports various privacy computing technologies and can be assembled flexibly to meet the needs of different scenarios.
Transparency: Build a unified technical framework, and try to make the underlying technology iteration transparent to the upper-layer application, with high cohesion and low coupling.
Openness: People with different professional directions can easily participate in the construction of the framework, and jointly accelerate the development of privacy computing technology.
Connectivity: Data in scenarios supported by different underlying technologies can be connected to each other.