Client Data Hashing for Fast LDP Accuracy Evaluation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems fail to efficiently evaluate and implement local differential privacy (LDP) protocols, leading to inefficiencies in determining the accuracy of LDP protocols, and existing systems require large amounts of computing time and resources to assess the accuracy of these protocols.
Innovation Solution
Implement a system where client devices perform local differential privacy (LDP) operations, including generating hash functions, adding noise to output vectors, and transmitting encrypted vectors to a recipient processing system, which updates a matrix with the data, allowing for more rigorous and computationally efficient evaluation of LDP protocol accuracy without performing the operations of each protocol.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional systems perform repeated operations of different LDP protocols to evaluate accuracy, then measurement precision of LDP protocol accuracy is improved, but loss of time and computing resources increases significantly
Solution Approach 1:
The patent pre-calculates and stores variance values for different LDP protocol parameters in a variance table before actual evaluation needs arise. When evaluating an LDP protocol, the system directly queries this pre-computed variance table rather than performing repeated operational simulations, thereby obtaining accurate precision measurements without the time cost of repeated experiments
Solution Approach 2:
The patent creates a variance table that copies and stores pre-computed variance data for various LDP protocol configurations. This copied variance information serves as a reference that eliminates the need to re-perform computational operations for evaluation, replacing expensive repeated simulations with efficient table lookups
2Adaptability or versatility
If conventional systems perform operations of multiple different LDP protocols to compare accuracy, then adaptability in evaluating different protocols is improved, but device complexity and computing resources increase
Solution Approach 1:
The patent creates a universal variance table that stores variance data applicable to multiple different LDP protocols and parameter configurations. This single variance table structure can serve evaluation needs for various protocols by accepting different protocol identifiers and parameter combinations as inputs, eliminating the need to implement separate evaluation systems for each protocol
Solution Approach 2:
The variance table acts as an intermediary data structure that mediates between different LDP protocols and the evaluation process. Instead of directly implementing and comparing multiple protocol operations, the system uses this intermediate variance table to store and retrieve pre-computed statistical properties, simplifying the comparison process across different protocols
3Reliability
If client devices perform local differential privacy operations with multiple hash functions and noise addition, then data security and privacy are improved, but computation cost at client devices increases
Solution Approach 1:
The patent allows the server to dynamically adjust LDP protocol parameters such as the number of hash functions, noise scale parameters, and probability distributions based on the evaluated variance data. By selecting protocols with optimal parameter configurations that achieve required privacy guarantees with minimal computational overhead, the system balances security requirements against client device computation costs
Data Source
AI summary
Methods, systems, and apparatus, including computer programs encoded on computer storage media, for evaluating a protocol for hashing data on client devices. One of the methods includes obtaining a size of a hashing domain for one or more hash functions, each configured to process source data to generate an output value; obtaining a probability of including each output value in a corresponding message; obtaining a differential privacy parameter; and generating a variance of the quantity of times the source data was a cause of a message over the one or more hash functions using the size of the hashing domain, the probability, and a value based on the differential privacy parameter.