Vertical Federated Random Forest Training With Encrypted Node Membership
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The security of sample data is low during the training of random forest models using vertical federated learning, as information is collected by a trusted third-party server, compromising data privacy.
Innovation Solution
Implement semi-homomorphic encryption and differential privacy processing to encrypt and process ciphertext calculation results in a ciphertext space, ensuring data security and privacy during vertical federated learning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If information such as feature data and intermediate values is collected by a trusted third-party server for vertical federated learning, then the training process can be completed, but the sample security is low and data privacy is compromised
Solution Approach 1:
The patent introduces semi-homomorphic encryption as an intermediary mechanism between participants and the third-party server. The encryption scheme allows the server to perform necessary computations on encrypted data without being able to decrypt or learn the underlying plaintext information, thus acting as a computational mediator that preserves privacy while enabling training completion
Solution Approach 2:
The patent transforms the data representation from plaintext to ciphertext through semi-homomorphic encryption. This parameter change in data state allows the system to maintain the functional requirements for model training while fundamentally altering the security properties of the data, making it unreadable to the server yet computable
2Productivity
If feature data and intermediate values are transmitted to the third-party server, then model training can proceed, but reverse inference attacks become possible
Solution Approach 1:
Semi-homomorphic encryption serves as a protective intermediary that allows the third-party server to receive and process encrypted feature data and intermediate values necessary for model training, while the mathematical properties of the encryption scheme prevent any reverse inference attacks by making the ciphertext computationally indistinguishable from random data
Solution Approach 2:
The patent replaces the traditional plaintext data transmission and processing mechanism with a cryptographic-based mechanism. Instead of relying on trust or security protocols, the system uses the mathematical foundations of semi-homomorphic encryption to provide inherent protection against reverse inference while maintaining training functionality
Data Source
AI summary
The disclosure relates to a method, apparatus, electronic device and readable storage medium for vertical federated learning. In the method, upon a plurality of participants training a random forest model based on a common sample participating in vertical federated learning, for a node to be trained on a decision tree included in the random forest model, the plurality of participants encrypt an identity configured to indicate whether the training sample is present in the node to be trained by a semi-homomorphic encryption technology, and then send it to the counterparty, and a further participant calculates a required ciphertext calculation result in a ciphertext space; in addition, the ciphertext calculation result is processed using a differential privacy technology and fed back to the counterparty, ensuring the security of the issued ciphertext calculation result, avoiding reversely interfering from the ciphertext calculation result to obtain related information of the sample, and improving the security of the sample.