Irreversible Feature Vector Transformation for Secure Data Exchange
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The challenge lies in securely exchanging data for machine learning tasks between providers and acquirers while protecting user privacy, as existing methods face difficulties in ensuring the legal use and preventing unauthorized reuse or leakage of plaintext data.
Innovation Solution
A method and device for transforming data into irreversible output feature vectors using a machine learning model, which are then combined with identifiers to create exchange samples that can be applied to machine learning, ensuring data security and privacy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If data is exchanged in plaintext for machine learning tasks, then data exchange efficiency is improved, but user privacy protection deteriorates and unauthorized reuse or leakage becomes easier
Solution Approach 1:
The patent changes the parameter of data representation from plaintext to encrypted form. The encryption algorithm transforms the data into a different parameter space where it remains usable for machine learning while being inaccessible in its original form, thus resolving the contradiction between exchange efficiency and privacy protection
Solution Approach 2:
The patent introduces an encryption mechanism as an intermediary between the data provider and acquirer. This intermediary layer allows data to be exchanged while maintaining security, as the encrypted data can be processed by the acquirer without revealing the original plaintext, thus solving the privacy leakage issue
2Object-affected harmful factors
If data is encrypted to protect user privacy, then privacy protection is improved, but data usability for machine learning deteriorates
Solution Approach 1:
The patent applies parameter changes by using encryption that preserves the structural properties of the data. The encrypted data maintains the necessary mathematical characteristics for machine learning operations while being protected, thus achieving both privacy protection and data usability simultaneously
Solution Approach 2:
The patent creates an encrypted copy of the data that serves the same functional purpose as the original for machine learning tasks. This encrypted copy can be processed and used for training models while the original plaintext remains protected, resolving the contradiction between privacy and usability
3Reliability
If data is transformed into feature vectors using machine learning models, then data security is improved through irreversible transformation, but processing complexity increases
Solution Approach 1:
The patent applies self-service by using the data itself to generate its own protective transformation. The machine learning model processes the data to create feature vectors that are inherently secure, as the transformation is based on the data's own characteristics and the model's learned patterns, thus achieving security without requiring external complex protection mechanisms
Data Source
AI summary
A data exchange method, a data exchange device, and a computing device for data exchange between a provider and a recipient for machine learning, the method including: (a) receiving a machine learning model from the provider (S1100); (b) respectively transforming output data samples into corresponding output eigenvectors by utilizing the machine learning model from the provider (S1200); (c) after transformation, combining the output eigenvectors with corresponding identifiers to form exchange samples (S1300). According to the data exchange method, original data is transformed into vector information which cannot be restored but can be applied to machine learning, for use in exchange, so as to, on one hand, enable efficient use of data for machine learning and, on the other hand, prevent unauthorized use, sale or disclosure of the original data.