Secure Multiparty Computation for Private Machine Learning

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

In distributed systems, institutions face challenges in analyzing data due to security concerns, leading to a loss of opportunity cost and limited potential for full data analysis.

Innovation Solution

An apparatus for secure multiparty computations using a processor and memory to submit secure multiparty computation requests onto an immutable sequential listing, involving contingent payments and authenticity commitments, and performing joint training protocols with localized models from participating devices.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If institutions make their data private for security reasons, then data security is improved, but data analysis capability deteriorates

Engineering Contradiction:

Improvedata securityVSAvoiddata analysis capability

Core Design Contradiction:

ReliabilityVSProductivity

Solution Approach 1:

The patent introduces secure multiparty computation protocols as an intermediary mechanism that enables joint model training across institutions without exposing private data. The computation request is submitted to an immutable sequential listing, and participating devices perform computations on encrypted data, allowing data analysis capability to improve while maintaining data security through cryptographic protection throughout the process.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent replaces traditional mechanical data sharing approaches (where data must be physically accessed or transferred) with cryptographic mechanisms. Instead of sharing plaintext data between institutions, the system uses authenticated encryption, digital signatures, and secure multi-party computation protocols to enable analysis while keeping data private, thus resolving the contradiction between security and analytical capability.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Productivity

If institutions share data for analysis, then data analysis capability is improved, but data privacy deteriorates

Engineering Contradiction:

Improvedata analysis capabilityVSAvoiddata privacy

Core Design Contradiction:

ProductivityVSLoss of information

Solution Approach 1:

The patent uses secure multiparty computation as an intermediary that allows institutions to collaborate on model training without directly sharing their private datasets. The system processes encrypted data through the immutable sequential listing and distributed computation nodes, extracting analytical value while preserving the confidentiality of individual institution's data through cryptographic safeguards.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent segments the data analysis process into distributed computations performed by multiple independent devices. Each institution's data remains localized and private, while the collective analytical capability is achieved through coordinated secure computations on segmented data portions, preventing any single institution from accessing other institutions' raw data.

Inventive Principle:

Principle #1Segmentation

3Productivity

If institutions collaborate on data analysis, then analytical potential is improved, but security risk deteriorates

Engineering Contradiction:

Improveanalytical potentialVSAvoidsecurity risk

Core Design Contradiction:

ProductivityVSObject-affected harmful factors

Solution Approach 1:

The patent introduces multiple intermediary security mechanisms including the immutable sequential listing, authenticated encryption layers, and secure multiparty computation protocols. These intermediaries enable collaborative analytical potential while mitigating security risks by ensuring that no single point of failure exists and that all data transmissions and computations are cryptographically protected throughout the collaborative process.

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS12278907B2Apparatus for secure multiparty computations for machine-learning

AI summary

An apparatus for secure multiparty computations for machine-learning is presented. The apparatus includes at least a processor and a memory communicatively connected to the at least a processor. The memory contains instructions configuring the at least a processor to submit a secure multiparty computation request onto an immutable sequential listing, wherein the secure multiparty computation request includes a contingent payment and an authenticity commitment of a first private dataset, receive at least a participant commitment from each participating device of a quorum of participating devices, generate a first localized model as a function of the first private dataset, and perform a joint training protocol as a function of the first localized model and a second localized model from the quorum of participating devices, wherein the joint training protocol includes generating a joint training datum.