ML Model Selection for Resilient Service Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional service processing systems require manual selection of machine learning models, which is costly and inflexible, and lack disaster recovery capabilities when models fail or network connections are lost.
Innovation Solution
Automatically select machine learning models based on model capability information and requirements, using a model filtering unit to determine suitable models from a set of machine learning models, and a model distributing unit to manage service requests efficiently.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If manual selection of machine learning models is used, then model selection can be performed, but it is costly and inflexible
Solution Approach 1:
The system automatically selects machine learning models based on capability matching between services and models, eliminating the need for manual model selection. The service processing system performs self-service by autonomously determining which models to invoke for given service requests, thereby reducing costs and improving flexibility.
Solution Approach 2:
The patent replaces the manual (mechanical) model selection process with an automated information processing system. The service processing system uses capability information databases and automatic matching algorithms to substitute human operators in the model selection task, achieving both cost reduction and increased automation.
2Reliability
If a single machine learning model is used for a service, then service processing is simple, but disaster recovery capability is lost when the model fails
Solution Approach 1:
The system performs preliminary actions by pre-establishing multiple machine learning models with different capability profiles before service requests arrive. When a service request is received, the system can select from multiple pre-available models, ensuring continuity of service even if one model fails. This advance preparation enables disaster recovery without adding operational complexity.
Solution Approach 2:
The patent changes the parameter of model selection from fixed single-model assignment to dynamic multi-model selection based on capability parameters. The service processing system evaluates capability information and selects appropriate models based on matching parameters, allowing flexible substitution of failed models with alternative models that meet the same service requirements.
3Productivity
If machine learning models are selected without capability evaluation, then model deployment is fast, but service processing efficiency decreases
Solution Approach 1:
The system performs capability evaluation of machine learning models in advance, before actual service processing. Capability information is pre-collected and stored in a database, allowing the service processing system to quickly match models to services without performing evaluation at request time. This preliminary action improves service processing efficiency while managing model capability information systematically.
Solution Approach 2:
The patent introduces an intermediary capability information database that mediates between machine learning models and service processing. This intermediary stores pre-evaluated capability information and enables efficient matching without direct complex evaluations during service execution, thereby improving productivity while maintaining organized capability management.
Data Source
AI summary
Embodiments of the disclosure provide a method, apparatus, device, storage medium, and program product for service processing. An example method includes: obtaining model capability information for a set of machine learning models, the model capability information comprising a respective evaluation result of each machine learning model in the set of machine learning models in a plurality of capability dimensions; based on a model capability requirement of a target service and the model capability information, selecting, from the set of machine learning models, at least one machine learning model satisfying the model capability requirement; and in response to receiving a service request for the target service, processing the service request with one or more machine learning models among the at least one machine learning model.