Network Model Association for Real-Time AI Model Management
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The management of AI models in network devices is challenging when multiple models are present and continuously trained and updated, as existing systems lack effective methods to associate and manage these models efficiently.
Innovation Solution
A network model management method and apparatus that involves obtaining and determining information about network models and analysis functions, allowing devices to establish association relationships between them for real-time management, including the use of provider, capability, and application scope information to facilitate the integration and updating of AI models.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If multiple AI models are continuously trained and updated in network devices, then the intelligence and functionality of network devices is improved, but the complexity of managing these models increases
Solution Approach 1:
The patent segments the AI model management process into distinct components: model information (first information) and function association information (second information). This segmentation allows the management system to handle multiple models systematically by processing and associating them with network analysis functions through structured data exchange between first device and second device
Solution Approach 2:
The patent introduces an intermediary mechanism where the first device obtains model information and determines associated functions, then sends this to the second device which establishes the actual association relationships. This intermediary process (information exchange and determination) simplifies the management complexity by distributing the management workload and using standardized information formats
2Reliability
If AI models are continuously trained and updated, then the model performance is improved, but the time required for management operations increases
Solution Approach 1:
The patent applies preliminary action by pre-establishing the information structure and association mechanisms before actual model deployment. The first device pre-obtains model information and pre-determines associated functions, creating a ready-to-execute association framework that reduces real-time management overhead when new models are deployed
Solution Approach 2:
The patent implements feedback mechanisms where the first device sends model information and function association data to the second device, which then establishes and maintains the association relationships. This feedback loop allows for efficient model registration and association without requiring manual intervention, reducing management operation time
3Adaptability or versatility
If network devices depend on external AI models, then the functionality is enhanced, but the ability to manage and control the models is reduced
Solution Approach 1:
The patent makes the management system universal by creating a standardized interface and information format that can handle multiple AI models from different providers. The first device can obtain information about various models and the second device can establish associations with different network analysis functions, making the management system adaptable to diverse model types and sources
Solution Approach 2:
The patent implements self-service by enabling the first device to automatically obtain model information, determine associated functions, and send this to the second device for association establishment. This automated self-service mechanism reduces the need for manual model management operations while maintaining full control capability
Data Source
AI summary
This application provides a network model management method applied to a first device. The method includes obtaining first information of a network model, determining second information based on the first information, and sending the first and second information to a second device. The first information describes the network model. The second information indicates a network analysis function, and the network analysis function is allowed to use the network model. The foregoing method is used, so that the second device can establish an association relationship between the network model and the network analysis function, to implement real-time management for the network model.