Client-Side Data Hashing With Noise for Private Transmission
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems face challenges in ensuring data privacy and security during data transmission over networks, particularly when client devices send unencrypted data to recipient processing systems, making them vulnerable to malicious access and data breaches.
Innovation Solution
Client devices perform local differential privacy operations using hash functions and noise introduction to encrypt and anonymize data before transmission, while recipient systems use hash functions to update matrices for accurate data processing without revealing sensitive information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If client devices transmit plain text data to recipient processing systems, then data communication is simple and fast, but data privacy and security are compromised
Solution Approach 1:
The client device performs differential privacy operations locally before data transmission, including generating hash values for data values and adding noise values. This preliminary processing ensures that sensitive information is anonymized and protected against malicious access before the data even leaves the client device, resolving the contradiction by preparing data in advance for secure transmission without compromising speed.
Solution Approach 2:
The patent introduces hash functions and noise values as intermediary elements between the original data and the transmitted data. The hash function transforms data values into hash values, and noise values are added to further anonymize the data. These intermediaries act as protective layers that maintain data privacy and security while allowing the data to be transmitted efficiently over the network.
2Object-affected harmful factors
If client devices perform local differential privacy operations using hash functions, then data privacy and security are improved, but communication cost increases
Solution Approach 1:
The patent changes the parameters of data representation by using hash functions that map data values to a condensed hash space. Instead of transmitting the full original data, only the hash values and noise values are transmitted. This parameter transformation reduces the amount of data that needs to be communicated while maintaining privacy and security, thereby reducing communication cost.
Solution Approach 2:
The patent extracts only the essential information needed for analysis by using hash functions to represent data values. The hash function captures the meaningful characteristics of the data while discarding unnecessary details. By extracting only the hashed representations and adding minimal noise, the system reduces communication overhead while preserving the necessary data for analysis.
3Object-affected harmful factors
If hash functions are used to anonymize data, then data privacy is enhanced, but data collisions occur reducing accuracy
Solution Approach 1:
The patent introduces noise values as an intermediary element that is added to the hash values before transmission. The noise values act as a perturbation layer that prevents direct mapping from transmitted data back to original values, thereby maintaining privacy. At the same time, the recipient can still retrieve accurate statistics by aggregating the noisy hash values, as the noise averages out over multiple samples.
Solution Approach 2:
The patent applies partial anonymization by using hash functions that provide privacy protection while maintaining sufficient accuracy for analysis. Instead of complete anonymization that would lose all information, the system uses hash functions that preserve the structural relationships and statistical properties of the data while providing privacy. This partial action approach balances privacy enhancement with measurement precision.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Methods, systems, and apparatus, including computer programs encoded on computer storage media, for hashing data on client devices. One of the methods includes generating, using a hash function, an output value that is in a hashing domain for the hash function; accessing a message size that is a smaller value than a domain size of the hashing domain; generating, using the hash function and the message size, one or more noise values; generating a message that includes the one or more noise values; and providing, to an external system, the message.