Encrypted Demand Projection Aggregation Without Data Sharing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Competing companies face challenges in forecasting market demand using historical data while preserving data privacy, as sharing data can compromise sensitive information and relying on a trusted third party introduces security risks and costs.
Innovation Solution
A cryptographic protocol using additive homomorphic encryption (AHE) allows companies to share encrypted partial projections of demand forecasts without revealing private data, utilizing a public key for encryption and a secret key for decryption to generate a combined plaintext forecast.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If companies share plaintext historical data for demand forecasting, then forecasting accuracy is improved, but data privacy and security are compromised
Solution Approach 1:
A trusted aggregator acts as an intermediary that collects encrypted partial projections from multiple companies, combines them, and returns the combined forecast without any company exposing their plaintext data. The aggregator handles the combined encrypted data through homomorphic operations, enabling accurate forecasting while maintaining data privacy through the intermediary mechanism.
Solution Approach 2:
Companies create encrypted copies (partial projections) of their historical data that can be shared and combined without revealing the original plaintext information. These encrypted copies contain the necessary information for forecasting when aggregated, but cannot be reverse-engineered to extract individual company data, thus enabling forecasting accuracy while preserving data privacy.
2Measurement precision
If companies use a trusted third party to combine data for forecasting, then forecasting accuracy is improved, but security risks and costs increase
Solution Approach 1:
The trusted aggregator serves as a mediator that only handles encrypted data throughout the process. It performs homomorphic addition on encrypted partial projections and returns encrypted combined results, never accessing plaintext data. This reduces security risks compared to traditional trusted third-party approaches because the aggregator cannot reverse-engineer individual company data from the encrypted combinations.
Solution Approach 2:
The patent replaces the mechanical system of trusted third-party data handling with cryptographic operations. Instead of relying on the trustworthiness of a third party to handle plaintext data securely, the system uses homomorphic encryption properties to perform computations on encrypted data, substituting cryptographic mechanisms for trust-based security models.
3Object-affected harmful factors
If companies share encrypted data, then data privacy is maintained, but the ability to perform accurate combined analysis is reduced
Solution Approach 1:
The patent substitutes traditional decryption-based analysis with homomorphic encryption operations. The trusted aggregator performs homomorphic addition on encrypted partial projections, which mathematically preserves the ability to compute accurate combined forecasts while maintaining encryption throughout the process. This cryptographic mechanism enables precise combined analysis without requiring decryption or compromising data privacy.
Data Source
AI summary
Aspects of the current subject matter are directed to privacy-preserving demand estimation. According to an aspect, a method includes receiving, by a first party, a request to provide a projection; receiving, by the first party, a key; in response to the request, performing, by the first party, a partial projection using private data of the first party without sharing the private data with at least one other party; encrypting, by the first party, the partial projection using the key; sending, by the first party, the encrypted partial projection to an aggregator; combining, by the aggregator; the encrypted partial projection provided by the first party with at least one other encrypted partial projection provided by the at least one other party to generate an encrypted combined projection. Related systems, methods, and articles of manufacture are also disclosed.


