Centralized AI Model Update via Distributed Logic Sharing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies face challenges in utilizing customer data with artificial intelligence (AI) or machine learning (ML) tools due to the need to safeguard sensitive customer information, which hinders the ability to enhance or optimize services based on customer interactions.
Innovation Solution
A system is developed that processes customer data to remove sensitive information, allowing for the sharing of abstracted logic patterns with a centralized AI/ML model. This decentralized approach enables individual entities to enhance services while maintaining data privacy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If customer data is processed by centralized AI/ML models to optimize services, then service optimization and efficiency improve, but customer data privacy and security deteriorate
Solution Approach 1:
The system segments the centralized model updating process into distributed contributions from multiple customer instances. Each instance independently contributes model logic updates without exposing raw customer data, dividing the data processing function across decentralized nodes while maintaining centralized model coordination.
Solution Approach 2:
The patent introduces an intermediary mechanism where model logic serves as a mediator between customer data and the centralized AI/ML model. Customer instances process their data locally to generate model logic contributions, which then update the centralized model without raw data leaving the customer instances, thus protecting privacy while enabling model improvement.
2Object-affected harmful factors
If sensitive customer information is retained and processed locally, then data privacy is maintained, but service optimization capability deteriorates
Solution Approach 1:
The system extracts only the essential model logic contributions from customer data processing while leaving the sensitive customer information retained locally at customer instances. This extraction process separates the useful learning signal from the sensitive data, allowing model improvement without data exposure.
Solution Approach 2:
The patent transforms customer data into model logic parameters through local processing at customer instances. This parameter transformation changes the data from sensitive raw customer information into abstracted model updates that preserve privacy while maintaining the ability to improve service optimization capabilities at the centralized level.
3Loss of time
If centralized models process all customer data, then learning efficiency improves, but system complexity and data security requirements deteriorate
Solution Approach 1:
The system segments the data processing workload from the centralized model, assigning local data processing to customer instances while reserving model coordination for the centralized system. This segmentation reduces the centralized model's data handling complexity and security burden while maintaining learning efficiency through coordinated updates.
Data Source
AI summary
Method and computer-readable media for updating a group ML model based on shared logic from customer data instances. The method includes receiving, at a communication interface of a central AI model, model logic from multiple remote customer instances of AI models, each customer instance of the AI models being based on the central AI model. The method includes updating the central AI model based on a combination of the model logic from the multiple remote customer instances. The method includes providing, via the communication interface, an AI model update to at least a subset of the multiple remote customer instances of the AI models.