CN-121981089-A - Data processing method and device

CN121981089ACN 121981089 ACN121981089 ACN 121981089ACN-121981089-A

Abstract

The application provides a data processing method and a data processing device, wherein the data processing method comprises the steps of obtaining a stem to be backfilled and a reference answer corresponding to the stem to be backfilled, inputting the stem to be backfilled and the reference answer into a question backfill model to obtain a target question output by the question backfill model, wherein the question backfill model determines the stem semantics of the stem to be backfilled, identifies at least one answer position in the stem to be backfilled, and backfills the reference answer to the stem to be backfilled based on the stem semantics and each answer position to generate the target question. By the data processing method provided by the application, the accuracy of backfilling is improved in the process of backfilling the answers to the stems.

Inventors

LIU HUAZHENG
GAO YIFEI
CHEN LEI
Weng Qiujie
LIU JINGMING

Assignees

北京猿力未来科技有限公司

Dates

Publication Date: 20260505
Application Date: 20260128

Claims (13)

1. A method of data processing, comprising: acquiring a stem to be backfilled and a reference answer corresponding to the stem to be backfilled; Inputting the stem to be backfilled and the reference answers to a stem backfill model to obtain a target question output by the stem backfill model, wherein the stem backfill model determines stem semantics of the stem to be backfilled, identifies at least one answer position in the stem to be backfilled, and backfills the reference answers to the stem to be backfilled based on the stem semantics and the answer positions to generate the target question.
2. The method of claim 1, wherein inputting the stem to be backfilled and the reference answer to a question backfill model to obtain a target question output by the question backfill model comprises: inputting the stem to be backfilled and the reference answer to a question backfill model; identifying the stem semantics corresponding to the stem to be backfilled based on the stem backfill model, and determining at least one answer position corresponding to the stem to be backfilled; And backfilling the reference answer to the stem to be filled based on the stem semantics corresponding to the stem to be backfilled and each answer position to obtain a target question output by the question backfill model.
3. The method of claim 1, wherein the topic backfill model is pre-trained by: obtaining a sampling stem and at least one sample reference answer corresponding to the sampling stem; Inputting the sample question stems and each sample reference answer to an initial question backfill model, and obtaining a predicted sample question output by the initial question backfill model, wherein the predicted sample question is generated by filling the sample reference answer into the sample question stems according to the sample question stem semantics and the sample answering positions, and the sample question stem semantics and the sample answering positions are determined according to the sample question stems; And adjusting model parameters corresponding to the initial question backfill model based on the predicted sample questions to obtain a question backfill model.
4. The method of claim 3, wherein inputting the sample stems and each sample reference answer to an initial question backfill model and obtaining a predicted sample question output by the initial question backfill model comprises: In the initial question backfill model, identifying sample question stem semantics corresponding to the sample question stems, and determining at least one sample answer position corresponding to the sample question stems; backfilling each sample reference answer to the sample stem based on the sample stem semantics and each sample answer position to generate and output a predicted sample question.
5. The method of claim 3, wherein obtaining a sample stem and at least one sample reference answer corresponding to the sample stem comprises: obtaining a sample question library, wherein the sample question library comprises at least one sample question stem and at least one sample reference answer; Determining sample stem semantics corresponding to each sample stem and sample reference answer semantics corresponding to each sample reference answer; based on the semantics of each sample stem and the semantics of each sample reference answer, determining a sample stem and at least one sample reference answer corresponding to the sample stem.
6. The method of claim 5, wherein determining the sample stem and at least one sample reference answer corresponding to the sample stem based on each sample stem semantic and each sample reference answer semantic comprises: Determining an initial sample stem and initial sample stem semantics corresponding to the initial sample stem, wherein the initial sample stem is any one of the sample stems; matching the initial sample stem semantics with each sample reference answer semantics to obtain at least one matching result; and determining a sample reference answer corresponding to the matching result meeting the matching condition as the sample reference answer corresponding to the initial sample stem.
7. The method of claim 3, wherein adjusting model parameters corresponding to the initial topic backfill model based on the predicted sample topic, to obtain a topic backfill model, comprises: determining a confidence score of the initial topic backfill model based on the predicted sample topic output, wherein the confidence score is used to characterize the accuracy of the predicted sample topic; And under the condition that the confidence score is smaller than a confidence score threshold, adjusting model parameters corresponding to the initial question backfill model until the confidence score of the question backfill model based on the predicted sample question output is larger than or equal to the confidence score threshold, so as to obtain the question backfill model.
8. The method of claim 1, wherein the answer locations included in the stem to be backfilled include at least one of filled-in underlines, brackets, option letter identification bits, text locations of semantically indicated answers.
9. The data processing method is applied to cloud side equipment and is characterized by comprising the following steps of: Receiving a stem to be backfilled and a reference answer corresponding to the stem to be backfilled, which are sent by a terminal side device; Inputting the stem to be backfilled and the reference answers to a stem backfill model to obtain a target question output by the stem backfill model, wherein the stem backfill model determines stem semantics of the stem to be backfilled, identifies at least one answer position in the stem to be backfilled, and backfills the reference answers to the stem to be backfilled based on the stem semantics and the answer positions to generate the target question; And sending the target title to end-side equipment.
10. A data processing apparatus, comprising: The device comprises an acquisition unit, a storage unit and a storage unit, wherein the acquisition unit is configured to acquire a stem to be backfilled and a reference answer corresponding to the stem to be backfilled; The processing unit is configured to input the stem to be backfilled and the reference answers to a stem backfilling model to obtain a target question output by the stem backfilling model, wherein the stem backfilling model determines stem semantics of the stem to be backfilled, identifies at least one answer position in the stem to be backfilled, and backfills the reference answers to the stem to be backfilled based on the stem semantics and the answer positions to generate the target question.
11. A computing device, comprising: A memory and a processor; the memory is adapted to store a computer program/instruction, the processor being adapted to execute the computer program/instruction, which when executed by the processor, implements the steps of the method according to any one of claims 1 to 9.
12. A computer readable storage medium storing a computer program/instruction, which when executed by a processor performs the steps of the method of any one of claims 1 to 9.
13. A computer program product comprising computer programs/instructions which, when executed by a processor, implement the steps of the method of any one of claims 1 to 9.

Description

Data processing method and device Technical Field The application relates to the technical field of computers, in particular to a data processing method. The application also relates to a data processing device, a computing device, a computer-readable storage medium and a computer program product. Background With the continuous development of computer technology, in the scenario of digital education question bank construction and intelligent automatic correction, in order to maintain the flexibility of the question bank, the stem and standard answer in the question bank are usually stored in an independent storage mode. However, when intelligent correction is performed, the standard answers need to be backfilled to the corresponding positions in the corresponding stems, so that stem answer-stem integrated data which can be used for intelligent correction is formed. In the process of backfilling standard answers in the question bank, the standard answers are usually matched according to preset matching rules, for example, by identifying fixed text identifiers in the question stems. However, this approach relies on the surface text features of the stem, lacking semantic understanding capabilities. Therefore, when there are various stem expressions, the backfill accuracy is low. Disclosure of Invention In view of this, the embodiment of the application provides a data processing method. The present application is also directed to a data processing apparatus, a computing device, a computer readable storage medium and a computer program product for solving the above-mentioned problems occurring in the prior art. According to a first aspect of an embodiment of the present application, there is provided a data processing method, including: acquiring a stem to be backfilled and a reference answer corresponding to the stem to be backfilled; Inputting the stem to be backfilled and the reference answers to a stem backfill model to obtain a target question output by the stem backfill model, wherein the stem backfill model determines stem semantics of the stem to be backfilled, identifies at least one answer position in the stem to be backfilled, and backfills the reference answers to the stem to be backfilled based on the stem semantics and the answer positions to generate the target question. According to a second aspect of the embodiment of the present application, there is provided a data processing method, applied to cloud-side equipment, including: Receiving a stem to be backfilled and a reference answer corresponding to the stem to be backfilled, which are sent by a terminal side device; Inputting the stem to be backfilled and the reference answers to a stem backfill model to obtain a target question output by the stem backfill model, wherein the stem backfill model determines stem semantics of the stem to be backfilled, identifies at least one answer position in the stem to be backfilled, and backfills the reference answers to the stem to be backfilled based on the stem semantics and the answer positions to generate the target question; And sending the target title to end-side equipment. According to a third aspect of an embodiment of the present application, there is provided a data processing apparatus including: The device comprises an acquisition unit, a storage unit and a storage unit, wherein the acquisition unit is configured to acquire a stem to be backfilled and a reference answer corresponding to the stem to be backfilled; The processing unit is configured to input the stem to be backfilled and the reference answers to a stem backfilling model to obtain a target question output by the stem backfilling model, wherein the stem backfilling model determines stem semantics of the stem to be backfilled, identifies at least one answer position in the stem to be backfilled, and backfills the reference answers to the stem to be backfilled based on the stem semantics and the answer positions to generate the target question. According to a fourth aspect of embodiments of the present application, there is provided a computing device comprising: A memory and a processor; The memory is used for storing computer programs/instructions, and the processor is used for executing the computer programs/instructions, and the computer programs/instructions realize the steps of the data processing method when being executed by the processor. According to a fifth aspect of embodiments of the present application, there is provided a computer readable storage medium storing a computer program/instruction which, when executed by a processor, implements the steps of the data processing method described above. According to a sixth aspect of embodiments of the present application, there is provided a computer program product comprising computer programs/instructions which when executed by a processor implement the steps of the data processing method described above. According to the data processing method provided by the application,