Ledger-Based AI Data Storing for Edge Device Security
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data sharing methods for artificial intelligence learning in edge computing lack secure, scalable, and verifiable platforms for transmitting and storing data across multiple parties, particularly in limited control environments.
Innovation Solution
A ledger-based AI data sharing system that uses distributed and collocated edge devices for secure, tamper-proof data transmission and storage, allowing for private and permissioned data flow between parties through a ledger node that generates and adds blocks to a ledger data structure containing AI model parameters and device health data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If distributed edge devices are used for AI learning, then data security and privacy are improved, but data sharing and verification across multiple parties become complex
Solution Approach 1:
The patent introduces a ledger node as an intermediary between edge devices and central systems. This mediator records AI model parameters and device health data in a blockchain ledger, enabling secure verification and sharing across multiple parties without requiring direct trust relationships. The intermediary resolves the complexity by providing a standardized, trusted interface for data exchange.
Solution Approach 2:
The system segments data into different types: raw edge device data remains local, processed AI model parameters are shared with the ledger node, and verification records are stored on the blockchain. This segmentation allows each party to maintain appropriate levels of data security while enabling necessary sharing for verification and collaboration.
2Measurement precision
If raw edge device data is transmitted for AI training, then model accuracy is improved, but data privacy and security are compromised
Solution Approach 1:
The patent extracts only the essential AI model parameters and device health data from the raw edge device data for storage on the ledger. By removing sensitive raw data and retaining only the processed, aggregated parameters needed for model improvement and verification, the system achieves both model accuracy and data privacy protection.
Solution Approach 2:
Instead of sharing raw data, the system creates and shares copies of processed AI model parameters on the blockchain ledger. These copies contain the learned information needed for model accuracy while being decoupled from the sensitive原始 data, thus protecting data privacy while maintaining model performance.
3Ease of operation
If centralized data storage is used for AI learning, then data access and model training are simplified, but security vulnerabilities and single points of failure increase
Solution Approach 1:
The patent segments the centralized storage architecture into a distributed ledger system where data is stored across multiple nodes. Each node holds a copy of the blockchain ledger, eliminating single points of failure while maintaining standardized data access through the consensus mechanism. This segmentation provides both security through distribution and ease of operation through standardized interfaces.
Solution Approach 2:
The system changes the fundamental parameter of data storage from centralized to distributed across the blockchain network. This parameter change transforms the security model from vulnerable centralized storage to resilient distributed storage, while the standardized blockchain protocol maintains ease of data access and model training operations.
4Speed
If data is processed and transmitted in real-time, then responsiveness is improved, but network bandwidth consumption and latency increase
Solution Approach 1:
The patent extracts and processes data locally at the edge devices, performing AI model training and parameter updates without transmitting raw data to central servers. Only the processed AI model parameters and device health data are transmitted to the ledger node, significantly reducing network bandwidth consumption while maintaining real-time processing capabilities at the edge.
Solution Approach 2:
The system performs preliminary data processing and AI model training at the edge devices before transmission. By completing the computationally intensive processing locally in advance, the system reduces the amount of data that needs to be transmitted over the network, thereby reducing bandwidth consumption and latency while maintaining responsiveness.
Data Source
AI summary
A system, apparatus, and method may include receiving edge device data, training a remote artificial intelligence (AI) model using the edge device data, and storing resulting remote AI model parameters in a ledger-based data structure. The ledger may also store device health data indicative of electronic device health of one or more edge devices. A ledger node may store a shared ledger. The shared ledger may use blockchain or encryption techniques and may store encrypted data, data pointers, or transfer data. The shared ledger may be accessible by at least two parties to share particular device health data.