Garbled Circuit Selection in Secure MPC With Secret Shares
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing secure multi-party computation (MPC) systems face challenges in efficiently selecting digital components for distribution to client devices while preserving user privacy and protecting data security, particularly in scenarios where multiple parties collaborate to make decisions based on user data without exposing the data in cleartext.
Innovation Solution
A garbled circuit protocol is employed within a secure MPC system involving multiple computers, using secret shares and oblivious transfer to select digital components based on user data conditions, ensuring that no individual party can access user data or intermediate values in cleartext, while minimizing data transmission and computational resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional secure multi-party computation protocols are used to select digital components while preserving user privacy, then user data security is improved, but computational costs and latency increase
Solution Approach 1:
The protocol segments user data into secret shares distributed across multiple parties, allowing computations to proceed on shared data without any single party having access to the complete original data. This enables secure component selection while maintaining computational efficiency by operating on smaller share representations rather than large encrypted datasets.
Solution Approach 2:
The protocol introduces garbled circuits as an intermediary mechanism that allows parties to compute selection results without directly exposing their secret shares. The garbled circuit acts as a trusted execution environment that processes conditions and user data shares to produce selection outcomes, reducing the computational burden on individual parties while maintaining security.
2Measurement precision
If comprehensive user data is processed to make accurate digital component selection decisions, then selection accuracy is improved, but data transmission bandwidth increases
Solution Approach 1:
The protocol extracts only the essential selection criteria and conditions from comprehensive user data, processing only the necessary attributes through garbled circuits. By selecting and processing only relevant data elements rather than complete user profiles, the system achieves accurate component selection with minimal bandwidth consumption.
Solution Approach 2:
The protocol applies different processing levels to different data elements, with sensitive attributes processed through secure garbled circuit evaluation while less sensitive attributes can be handled more efficiently. This localized approach to data processing quality allows accurate selection decisions without uniformly high bandwidth consumption across all data types.
3Reliability
If multiple parties collaboratively evaluate digital component eligibility conditions, then decision reliability is improved, but system complexity increases
Solution Approach 1:
The protocol merges the evaluation of multiple eligibility conditions into a unified garbled circuit structure, allowing parties to jointly assess user data against all conditions in a single coordinated process. This consolidation maintains decision reliability through multi-party verification while reducing protocol complexity compared to separate evaluation steps for each condition.
Solution Approach 2:
The garbled circuit mechanism serves multiple functions simultaneously: it protects secret shares, evaluates eligibility conditions, and produces selection results. This multi-functionality allows the same protocol structure to handle various eligibility criteria and data types, reducing overall system complexity despite the collaborative nature of the computation.
Data Source
AI summary
This document describes systems and techniques for using cryptography, secret sharing, secure MPC, garbled circuits, and oblivious transfer to select digital components in ways that preserve user privacy and protects the security of data of each party that is involved in the selection process. A method includes obtaining, by a first computer of a secure multi-party computation (MPC) system, at least a first share of user data related to a user of a client device. For each digital component in a set, a first secret share of a condition bit for the condition is obtained for each of one or more conditions that condition eligibility of the digital component for distribution. A garbled circuit is executed to select a given digital component for which each of the one or more conditions of the digital component is satisfied using the first secret share and the second secret share for each condition.