Biogenesis quality auxiliary evaluation method and device based on multi-modal large model

By processing student data using a multimodal large model, the problem of low efficiency in processing unstructured data in the evaluation of student quality in universities has been solved, and automated evaluation of both structured and unstructured data has been achieved, thus improving evaluation efficiency.

CN122364876APending Publication Date: 2026-07-10BEIJING UNIV OF POSTS & TELECOMM
0 Cites 0 Cited by

Patent Information

Application Number
CN202610663712.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-05-14
Publication Date
2026-07-10

AI Technical Summary

Technical Problem

The existing evaluation system for the quality of college students relies on structured data and lacks the ability to automatically process unstructured data, resulting in low evaluation efficiency.

Method used

Multimodal large model is used to process student data. The data is transformed into structured data through mapping rules. Attention mechanism and visual encoder are used to extract student quality features from unstructured text and image data. The data is then used for automated evaluation by combining preset prompt words and evaluation indicators.

Benefits of technology

It enables integrated processing of structured and unstructured data, improving the efficiency of student quality assessment, automating the processing of unstructured text and image materials, and reducing human intervention.

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

Abstract

The invention provides a biogenic source quality auxiliary evaluation method and device based on a multi-modal large model, and belongs to the technical field of large model evaluation, and the method comprises the steps: obtaining biogenic source data of a target person; performing feature extraction on the biogenic data by using a multi-modal large model based on the following mode to obtain biogenic quality features of the target person in multiple dimensions: converting structured data in the biogenic data into corresponding biogenic quality features according to a preset mapping rule; extracting biosource quality features corresponding to the unstructured text data in the biosource data based on a preset cue word by using an attention mechanism; a visual encoder is utilized to extract source quality features corresponding to the unstructured image data in the source data; the multi-dimensional biogenic quality features are used for assisting biogenic quality evaluation of the target personnel. According to the method, the structured data and the unstructured data in the biological source data can be processed at the same time, related features are extracted from the data, related personnel can conveniently carry out work, and the evaluation efficiency is improved.
Need to check novelty before this filing date? Find Prior Art