CN-121980136-A - Multi-mode talent assessment data intelligent analysis and safe operation and maintenance system and method based on large model

CN121980136ACN 121980136 ACN121980136 ACN 121980136ACN-121980136-A

Abstract

The application relates to the technical field of talent assessment, in particular to a multi-mode talent assessment data intelligent analysis and safety operation and maintenance system and method based on a large model, comprising a multi-mode data acquisition layer, a data analysis layer and a data analysis layer, wherein the multi-mode data acquisition layer is provided with sensing equipment and an interaction terminal and is used for acquiring multi-mode original data; the system comprises a data preprocessing layer, a large model analysis layer and a result output layer, wherein the data preprocessing layer is used for performing standardized conversion, noise filtering and feature extraction processing on multi-mode original data to obtain structural feature data, the large model analysis layer is deployed with a pre-training multi-mode large model and is used for performing dimension measurement reasoning based on the structural feature data and outputting an evaluation analysis result, and the result output layer is used for outputting the evaluation analysis result in a visual form or a structural report form. The data preprocessing layer is utilized to effectively process the data, intelligent and accurate evaluation analysis is realized by means of the large model analysis layer, a user can conveniently check an evaluation result through the result output layer, and meanwhile, the safety of the system in the whole operation process is guaranteed through the safety operation and maintenance layer.

Inventors

Guo nanming
MA HE
NI XIAOMING
DU YULIN
Hong Qiankai
Cui Haosong
JIANG YIXUAN

Assignees

网才科技(广州)集团股份有限公司

Dates

Publication Date: 20260505
Application Date: 20260122

Claims (10)

1. The utility model provides a multimode talent evaluation data intelligence analysis and safe fortune dimension system based on big model which characterized in that includes: The multi-mode data acquisition layer is configured with at least one sensing device and an interaction terminal and is used for acquiring multi-mode original data of an evaluation object, wherein the multi-mode original data comprises one or more of text data, voice data, image data and physiological characteristic data; The data preprocessing layer is in communication connection with the multi-mode data acquisition layer and is used for performing standardized conversion, noise filtering and feature extraction processing on the multi-mode original data to obtain structural feature data; the large model analysis layer is deployed with a pre-training multi-mode large model and is in communication connection with the data preprocessing layer and is used for executing evaluation dimension reasoning based on the structural characteristic data and outputting an evaluation analysis result; and the result output layer is in communication connection with the large model analysis layer and is used for outputting the evaluation analysis result to user equipment in a visual form or a structured report form.
2. The system of claim 1, wherein the sensing device of the multi-modal data collection layer comprises one or more of an image collection device, an audio collection device, and a physiological sensing device, and the interactive terminal comprises one or more of a computer terminal, a mobile terminal, and a dedicated evaluation terminal.
3. The system of claim 1, wherein the pre-training multi-mode large model is constructed based on a deep learning architecture and comprises a mode coding module and a cross-mode fusion module, wherein the cross-mode fusion module realizes fusion of different mode characteristics through an attention mechanism or a characteristic splicing mode, and training data of the pre-training multi-mode large model comprises a public assessment corpus and history assessment data after desensitization processing.
4. The system of claim 1, further comprising a security operation and maintenance layer in communication with the multi-modal data collection layer, the data preprocessing layer, and the large model resolution layer, respectively, the security operation and maintenance layer comprising a data encryption module, an access control module, and an anomaly monitoring module, The data encryption module is used for carrying out encryption protection on the full life cycle of the data; The access control module is used for controlling the data access range based on the authority allocation rule; the abnormality monitoring module is used for monitoring the operation behavior of the system in real time and triggering an alarm.
5. The system of claim 4, wherein the access control module supports a role-based rights allocation mechanism and configures a multi-role rights separation policy.
6. The system of claim 4, wherein the data encryption module supports data hierarchical encryption, wherein the data hierarchical encryption is implemented by adopting a corresponding encryption mode according to a data sensitivity level, and an encryption key adopted by encryption is dynamically generated and periodically updated through a key management system.
7. The system of claim 1, wherein the result output layer is configured with a standardized data interface that supports data interfacing with external business systems and employs secure communication protocols and passes identity authentication mechanisms.
8. A multi-mode talent assessment data intelligent analysis and safe operation and maintenance method based on a large model is characterized by comprising the following steps: S1, acquiring multi-mode original data of an evaluation object through sensing equipment and an interactive terminal of a multi-mode data acquisition layer, wherein the multi-mode original data comprises one or more of text data, voice data, image data and physiological characteristic data; S2, carrying out standardized conversion, noise filtering and feature extraction processing on the multi-mode original data through a data preprocessing layer to obtain structured feature data; s3, invoking a pre-trained multi-mode large model through a large model analysis layer, inputting the structural feature data, and outputting an evaluation analysis result through cross-mode feature fusion and reasoning; and S4, converting the evaluation analysis result into a visual form or a structural report through a result output layer, and pushing the visual form or the structural report to user equipment.
9. The method of claim 8, wherein the feature extraction algorithm in step S2 comprises one or more of a semantic feature extraction algorithm, an emotion feature extraction algorithm, an image feature extraction algorithm, and a physiological feature extraction algorithm.
10. The method of claim 8, wherein the large model resolution layer in step S3 further comprises a model optimization module that performs parameter adjustment on the pre-trained multi-modal large model using a machine learning optimization strategy based on the evaluation resolution result.

Description

Multi-mode talent assessment data intelligent analysis and safe operation and maintenance system and method based on large model Technical Field The application relates to the technical field of talent assessment, in particular to a multi-mode talent assessment data intelligent analysis and safe operation and maintenance system and method based on a large model. Background In the traditional talent assessment process, data with a single dimension are often relied on, such as talents are assessed only through pen test achievements or interview performances, and comprehensive quality of the talents is difficult to comprehensively and accurately reflect by the assessment mode. Along with the development of science and technology, multi-mode data are gradually introduced into talent assessment fields, wherein the multi-mode data comprise various types of data such as texts, voices, images and physiological characteristics, and talents can be described from various angles. However, how to efficiently and accurately collect the multi-mode data and effectively analyze the collected data, and simultaneously ensure the safety of the data in the whole process, becomes a problem to be solved in the current talent assessment field. Disclosure of Invention The application provides a multi-mode talent assessment data intelligent analysis and safety operation and maintenance method and system based on a large model, which realize mutual collaboration among a multi-mode data acquisition layer, a data preprocessing layer, a large model analysis layer, a result output layer and a safety operation and maintenance layer, and jointly complete intelligent analysis and safety operation and maintenance tasks of talent assessment data. In order to achieve the above purpose, the application adopts the following technical scheme: In a first aspect, a system for intelligent analysis and safe operation and maintenance of multi-mode talent assessment data based on a large model is provided, including: The multi-mode data acquisition layer is configured with at least one sensing device and an interaction terminal and is used for acquiring multi-mode original data of an evaluation object, wherein the multi-mode original data comprises one or more of text data, voice data, image data and physiological characteristic data; The data preprocessing layer is in communication connection with the multi-mode data acquisition layer and is used for performing standardized conversion, noise filtering and feature extraction processing on the multi-mode original data to obtain structural feature data; the large model analysis layer is deployed with a pre-training multi-mode large model and is in communication connection with the data preprocessing layer and is used for executing evaluation dimension reasoning based on the structural characteristic data and outputting an evaluation analysis result; and the result output layer is in communication connection with the large model analysis layer and is used for outputting the evaluation analysis result to user equipment in a visual form or a structured report form. In a second aspect, a method for intelligent analysis and safe operation and maintenance of multi-mode talent assessment data based on a large model is provided, which comprises the following steps: S1, acquiring multi-mode original data of an evaluation object through sensing equipment and an interactive terminal of a multi-mode data acquisition layer, wherein the multi-mode original data comprises one or more of text data, voice data, image data and physiological characteristic data; S2, carrying out standardized conversion, noise filtering and feature extraction processing on the multi-mode original data through a data preprocessing layer to obtain structured feature data; s3, invoking a pre-trained multi-mode large model through a large model analysis layer, inputting the structural feature data, and outputting an evaluation analysis result through cross-mode feature fusion and reasoning; and S4, converting the evaluation analysis result into a visual form or a structural report through a result output layer, and pushing the visual form or the structural report to user equipment. The system comprehensively collects multi-dimensional data of an evaluation object through the multi-mode data collection layer, effectively processes the data through the data preprocessing layer, achieves intelligent and accurate evaluation analysis through the large model analysis layer, facilitates a user to check an evaluation result through the result output layer, and simultaneously guarantees safety of the system in the whole operation process through the safety operation and maintenance layer. Compared with the traditional talent assessment method, the system has the remarkable advantages that the assessment object can be comprehensively assessed from multiple dimensions such as texts, voices, images and physiological characteristics, the limitation of singl