Encrypted ML Model Update Path for Trusted-Untrusted Isolation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies lack secure methods for updating machine learning models and do not effectively rate or request updates for stale models.
Innovation Solution
A system involving paired encryption devices with encryption units and communication units is used to securely update machine learning models by encrypting and decrypting packets with connectionless headers, ensuring secure communication across untrusted networks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing techniques are used for updating machine learning models, then the updating process is simple, but the security of receiving updates is insufficient
Solution Approach 1:
The patent introduces an encryption device as an intermediary component between the processing device and the update source. This encryption device includes an encryption unit that encrypts outgoing requests and a decryption unit that decrypts incoming updates using cryptographic keys, thereby securing the update transmission without requiring complex security infrastructure throughout the entire system.
Solution Approach 2:
The system is segmented into distinct functional components: a processing device that generates update requests, an encryption device that handles cryptographic operations, and a model aggregator that provides updates. This segmentation allows security to be implemented at specific critical points (in the encryption device) rather than requiring comprehensive security across all system components.
2Adaptability or versatility
If existing techniques are used for machine learning model updates, then the system operation is simple, but the ability to rate models and request updates for stale models is insufficient
Solution Approach 1:
The processing device is configured to rate the machine learning model and determine whether it is stale based on performance metrics. When the model is rated as stale, the system automatically generates an update request. This feedback mechanism enables adaptive model management without requiring manual intervention or complex external rating systems.
Solution Approach 2:
The system performs self-service by automatically rating its own machine learning model, determining staleness based on performance degradation, and requesting updates when needed. The processing device monitors model performance and autonomously initiates update requests, reducing the need for external model management infrastructure.
Data Source
AI summary
A system for facilitating secure updating of a machine learning model. The system includes a processing device and a first encryption device. The processing device generates a request and updates the machine learning model using an update. The first encryption device includes a first encryption unit and a first communication unit. The first encryption unit encrypts a native packet corresponding to the request and adds a connectionless header forming a first egressing connectionless datagram, decrypts a second encrypted native packet of an ingressing connectionless datagram to obtain the update. The first communication unit adds a complex header to the first egressing connectionless datagram for forming a first packet for delivery to a second encryption device, receives a second packet comprising the second encrypted native packet and a complex header from the second encryption device, removes the complex header and adds a connectionless header for forming the ingressing connectionless datagram