Expert knowledge-driven large-model customized data processing method and related equipment

Through the customized data processing method of large models driven by expert knowledge, and the knowledge embedded vector customization adapter expert submodel is used to solve the problem of insufficient accuracy and task conflicts of generalist large models on specific tasks, and realize the efficient and professional data processing of large models in multiple scenarios.

CN119670807AActive Publication Date: 2025-03-21BEIJING UNIV OF POSTS & TELECOMM
5 Cites 3 Cited by

Patent Information

Application Number
CN202510194176.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-02-21
Publication Date
2025-03-21
Estimated Expiration
2045-02-21

AI Technical Summary

Technical Problem

The existing generalist models have less accuracy than dedicated models on specific tasks, and semantic conflicts between tasks during multi-task hybrid training lead to negative performance migration.

Method used

A method of customized data processing of large-scale knowledge driven by expert knowledge is proposed. By inputting business data into a trained large-scale model that can dynamically customize expert sub-model, determining the weight of adapter experts based on the knowledge embedding vector, and customizing the expert sub-model to achieve efficient processing of business data.

Benefits of technology

On the basis of taking into account the generalization capabilities of large models and dedicated capabilities, a large model solves multiple services, has business availability in multiple scenarios, and improves the professionalism and adaptability of data processing.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention provides an expert knowledge-driven large-model customized data processing method and related equipment. The method comprises the following steps: inputting acquired business data into a trained large model capable of dynamically customizing an expert sub-model; in the large model, based on the business data, determining a knowledge embedding vector; based on the knowledge embedding vector, determining the weight of each adapter expert, and customizing an expert sub-model; and processing the business data based on the expert sub-model. According to the embodiment of the invention, the business knowledge hierarchical structure is constructed by using the multi-dimensional business information and the expert knowledge based on the large-scale business data. In combination with a large-model efficient fine tuning technology and a mixed expert strategy, dynamic expert group adaptation is realized under the guidance of business data knowledge, so that different expert models are customized for different business data, automatic adaptation under multiple scenes is realized, and more professional data processing is realized.
Need to check novelty before this filing date? Find Prior Art