An examination method and system combining a language model and virtual technology

Abstract

The application discloses a kind of examination method and system combined with language model and virtual technology, method includes the following steps: police and citizen dialogue process is rehearsed, and the first language model is used to analyze the rehearsed language output of subject, obtains rehearsal analysis result;Based on rehearsal analysis result, obtain the rehearsal score of subject;When the rehearsal score of subject meets preset score requirement, obtain examination permission;Based on examination permission, combined with virtual technology, build police and citizen dialogue scene topic;Based on police and citizen dialogue scene topic, examine subject, obtain examination language output;Second language model is used to analyze examination language output, obtain the examination evaluation report of subject.The application will be combined with language model and virtual technology, build more advanced, efficient examination method and system, improve the quality and effect of examination.

Description

Technical Field

[0001] This invention belongs to the field of virtual reality technology, specifically relating to an assessment method and system that combines language models and virtual technology. Background Technology

[0002] Virtual reality (VR) technology, as a cutting-edge development direction of the next generation of information technology, is gradually permeating various industries and fields. These technologies provide users with entirely new interactive experiences by simulating real-world environments or augmenting reality scenarios. With the improvement of computer graphics processing capabilities and the continuous advancement of display technology, the realism and immersion of virtual environments have been significantly enhanced, enabling virtual technology to demonstrate enormous application potential in multiple fields such as education, training, and entertainment.

[0003] Traditional assessment methods are often limited by factors such as venue, equipment, and resources, making it difficult to comprehensively and realistically simulate actual work or learning scenarios. Virtual technology, however, can overcome these limitations, providing a highly realistic virtual environment for assessment, thereby more accurately evaluating the abilities and qualities of test takers. Furthermore, traditional assessment methods often require manual scoring and evaluation, which suffers from high subjectivity and low efficiency. AI technology, on the other hand, can automate assessments, avoiding interference from human factors and improving the fairness and accuracy of evaluations. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides an assessment method and system that combines language models and virtual technology. By creating virtual reality scenarios, it assesses the comprehensive qualities of test takers and improves their police work capabilities.

[0005] To achieve the above objectives, the present invention provides the following solution:

[0006] An assessment method combining language models and virtual technology includes the following steps:

[0007] The drill involved a dialogue between police and civilians, and the first language model was used to analyze the language output of the participants to obtain the analysis results.

[0008] Based on the analysis results of the exercise, the exercise score of the subjects was obtained;

[0009] When the subject's practice score meets the preset score requirement, the subject is granted permission to perform the assessment.

[0010] Based on the aforementioned assessment and authorization, and combined with virtual technology, a scenario-based question for police-citizen dialogue is constructed.

[0011] The subjects were assessed based on the police-civilian dialogue scenario questions, and their assessment language output was obtained.

[0012] The assessment language output is analyzed using a second language model to obtain the assessment evaluation report of the subjects.

[0013] Preferably, the method for obtaining the exercise analysis results is as follows:

[0014] Collect problems from daily police work by category and establish a problem database;

[0015] The subjects were asked to conduct a police-civilian dialogue exercise based on the aforementioned question database, and the subjects' exercise language output was obtained.

[0016] Using a tree-structured long short-term memory network, the syntactic structure information of the exercise language output is extracted, and the syntactic structure information labels are obtained;

[0017] The syntactic structure information is input into the first BERT language model to obtain the sentiment classification probability value and keyword information;

[0018] Based on the syntactic structure information labels, the sentiment classification probability values, and the keyword information, a loss function for the syntactic structure information is obtained.

[0019] The loss value of the syntactic structure information is calculated using a loss function, and the parameters of the first BERT language model are optimized based on the loss value to obtain the optimal first BERT language model;

[0020] The exercise analysis results are obtained based on the optimal first BERT language model.

[0021] Preferably, the exercise analysis results include the results of the subjects' intention recognition ability analysis, the results of communication clarity analysis, the results of communication logical coherence analysis, and the results of the accuracy of regulatory application analysis.

[0022] The preferred method for constructing police-civilian dialogue scenario questions is as follows:

[0023] Test questions are extracted from the test question bank to obtain the test questions;

[0024] Create a 3D model of the scene and a 3D model of the virtual character that match the assessment questions;

[0025] The 3D model of the scene and the 3D model of the virtual character are imported into the virtual reality development platform for model assembly to obtain a virtual scene of police-civilian dialogue.

[0026] Based on the assessment questions, establish behavioral logic for the 3D model of the virtual character, and set triggers and interaction points;

[0027] The virtual scenario of police-civilian dialogue and the 3D model of the virtual character after establishing behavioral logic are deployed to a VR device to complete the construction of the police-civilian dialogue scenario.

[0028] Preferably, the method for analyzing the assessment language output using a second language model is as follows:

[0029] Based on the output of the assessment language, the reaction time of the subjects is collected, and the reaction characteristics of the subjects are extracted.

[0030] Based on the second BERT language model, syntactic structure analysis and keyword extraction are performed on the assessment language output to extract the language features of the subjects.

[0031] The subjects were assessed on their overall competence by combining the aforementioned response characteristics and language characteristics.

[0032] Based on the comprehensive quality assessment results, the subjects were classified into different levels and their deficiencies in police work ability were analyzed.

[0033] Based on the aforementioned grading and the deficiencies in police work capabilities, an assessment report for the subjects was obtained.

[0034] The present invention also provides an assessment system that combines language models and virtual technology. The method is characterized by comprising: a practice module, a practice scoring module, an assessment authorization module, an assessment module, and an assessment analysis module.

[0035] The training module is used to practice police-civilian dialogue and uses a first language model to analyze the language output of the subjects during the training to obtain training analysis results.

[0036] The exercise scoring module is used to obtain the subject's exercise score based on the exercise analysis results;

[0037] The assessment permission module is used to obtain assessment permission when the subject's practice score meets the preset score requirements;

[0038] The assessment module is used to construct police-civilian dialogue scenario questions based on the assessment permission and in conjunction with virtual technology; to assess the subjects based on the police-civilian dialogue scenario questions and obtain assessment language output;

[0039] The assessment analysis module is used to analyze the assessment language output using a second language model to obtain the assessment evaluation report of the subject.

[0040] Preferably, the training module includes a training language output acquisition unit, a syntactic information extraction unit, a syntactic information analysis unit, a loss function calculation unit, and a model optimization unit;

[0041] The language output acquisition unit is used to classify and collect daily police work issues and establish a problem database; it enables subjects to conduct police-civilian dialogue drills based on the problem database and obtains the subjects' drill language output.

[0042] The syntactic information extraction unit is used to extract the syntactic structure information of the exercise language output using a tree-structured long short-term memory network, and to obtain the syntactic structure information label.

[0043] The syntactic information analysis unit is used to input the syntactic structure information into the first BERT language model to obtain the sentiment classification probability value and keyword information;

[0044] The loss function calculation unit is used to obtain the loss function of the syntactic structure information based on the syntactic structure information label, the sentiment classification probability value, and the keyword information;

[0045] The model optimization unit is used to calculate the loss value of the syntactic structure information using a loss function, and optimize the parameters of the first BERT language model based on the loss value to obtain the optimal first BERT language model; and obtain the exercise analysis results based on the optimal first BERT language model.

[0046] Preferably, the assessment module includes: a question extraction unit, a 3D model creation unit, a model assembly unit, a behavioral logic design unit, and a deployment unit;

[0047] The question extraction unit is used to extract questions from the assessment question bank to obtain assessment questions.

[0048] The 3D model creation unit is used to create a scene 3D model and a virtual character 3D model that conform to the assessment questions.

[0049] The model assembly unit is used to import the scene 3D model and the virtual character 3D model into the virtual reality development platform to assemble the model and obtain a virtual scene of police-civilian dialogue.

[0050] The behavioral logic design unit is used to establish behavioral logic for the three-dimensional model of the virtual character based on the assessment questions, and to set triggers and interaction points.

[0051] The deployment unit is used to deploy the virtual scene of police-civilian dialogue and the 3D model of the virtual character after establishing behavioral logic to the VR device, thereby completing the construction of the police-civilian dialogue scene questions.

[0052] Compared with existing technologies, the beneficial effects of this invention are as follows: Through virtual reality (VR) technology, subjects can interact with the virtual environment in real time. This interaction method is more vivid and intuitive than traditional paper-and-pencil tests or oral defenses, helping to more comprehensively assess the subjects' overall qualities. Artificial intelligence (AI) technology can automatically collect and analyze data during the assessment process, quickly generating assessment reports and greatly improving assessment efficiency. At the same time, the AI ​​assessment process is unaffected by human factors, ensuring the fairness of the assessment. The combination of virtual technology and AI makes the design of assessment content more flexible and diverse. The difficulty, content, and format of the assessment can be quickly adjusted according to actual needs to meet the assessment requirements of different levels and needs. This invention also utilizes language models to analyze the language output of drills and assessments, objectively evaluating the subjects' overall qualities. Attached Figure Description

[0053] To more clearly illustrate the technical solution of the present invention, the drawings used in the embodiments are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0054] Figure 1 This is a flowchart of an assessment method combining language models and virtual technology according to an embodiment of the present invention. Detailed Implementation

[0055] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0056] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0057] Example 1

[0058] like Figure 1 As shown, an assessment method combining language models and virtual technology includes the following steps:

[0059] S1: The police-civilian dialogue process was rehearsed, and the first language model was used to analyze the language output of the subjects in the rehearsal to obtain the rehearsal analysis results;

[0060] A further implementation method is to obtain the exercise analysis results as follows:

[0061] Collect problems from daily police work by category and establish a problem database;

[0062] Subjects practiced police-civilian dialogues using a question database, generating their practiced language output. A tree-structured long short-term memory (LSTM) network was used to extract the syntactic structure information from the practiced language output and obtain syntactic structure information labels. In this embodiment, the syntactic structure information labels include hard labels and soft labels. Hard labels represent the true sentiment label of the sentence, and soft labels represent the sentiment probability distribution of the sentence. Specifically, the tree-structured long short-term memory network was used to learn the syntactic structure information of the existing dataset, encoding the sentence component structure tree. Next, a Stanford Parser was used to parse the practiced language output into a syntactic tree structure. Then, a tree-structured long short-term memory network with frozen parameters was used to compute the hidden representation of each node in the syntactic tree and obtain the soft label of each node, resulting in "text-soft label" samples. These training samples (practice language output) thus possess syntactic structure.

[0063] Syntactic structure information is input into the first BERT language model to obtain sentiment classification probability values ​​and keyword information; specifically, preset identifiers are added before and after the training samples and input into the first BERT language model, the output of the language model is encoded and input into the Softmax classifier to obtain the sentiment classification probability values ​​(hard labels) of the training samples.

[0064] Based on syntactic structure information labels (soft labels), sentiment classification probability values ​​(hard labels), and keyword information, a loss function for syntactic structure information is obtained. In this embodiment, for samples with soft labels, a weighted cross-entropy loss is used to calculate the soft label loss term; for samples with hard labels, a weighted cross-entropy loss is also used to calculate the hard label loss term; a pairwise ranking loss function is used to calculate the keyword loss term; and using preset loss weights, the soft label loss term, hard label loss term, and keyword loss term are weighted and combined to obtain the final loss function.

[0065] The loss function is used to calculate the loss value of syntactic structure information, and the parameters of the first BERT language model are optimized based on the loss value to obtain the optimal first BERT language model. Based on the final loss function, the total loss value is obtained, and the parameters of the first BERT language model are adjusted through a backpropagation neural network to improve the accuracy of the language model for sentiment classification. In this embodiment, the core of BERT is composed of multiple stacked Transformer encoder layers. Each encoder layer contains the following key components:

[0066] Multi-head self-attention mechanism: allows the model to focus on tokens at different positions when processing a sequence and calculate attention weights between tokens, thereby capturing the dependencies in the input sequence.

[0067] Feedforward neural networks: further transform the output of the self-attention mechanism to extract higher-level features.

[0068] Residual connections and layer normalization: used to improve the training stability and performance of the model, and help alleviate the problems of vanishing and exploding gradients.

[0069] Based on the optimal first BERT language model, the results of the exercise analysis were obtained.

[0070] Examples of emotion classification for test takers:

[0071] (1) In an emergency, how to calm down a person who is emotionally agitated due to a family dispute?

[0072] (2) How to clearly and patiently explain the relevant legal provisions to a client who does not understand legal procedures?

[0073] (3) When handling a traffic accident, how can we effectively communicate with both parties involved, collect key information, and maintain order at the scene?

[0074] By analyzing the subjects' rehearsal language output, we can obtain the subjects' emotional categories (positive, negative, etc.) in different rehearsal questions and whether the keywords were used incorrectly.

[0075] A further implementation method involves that the exercise analysis results include the results of the subjects' ability to recognize intent, the results of the clarity of communication, the results of the coherence of communication logic, and the results of the accuracy of the application of regulations.

[0076] S2: Based on the exercise analysis results, obtain the exercise score of the subject; score the subject for each exercise analysis category to obtain the exercise score.

[0077] S3: When the subject's practice score meets the preset score requirement, the subject is granted assessment permission; the conditions for obtaining assessment permission also include the total practice time meeting the preset practice time.

[0078] S4: Based on the assessment and approval process, and combined with virtual technology, construct scenarios and questions for police-citizen dialogue;

[0079] A further implementation involves using virtual technology to simulate various complex police situations and public reactions during police-community dialogue assessments, thereby more accurately evaluating the participants' response capabilities. The method for constructing police-community dialogue scenario questions is as follows:

[0080] Test questions are extracted from the test question bank to obtain the test questions;

[0081] Create 3D scene models and 3D virtual character models that match the assessment questions;

[0082] Import the 3D model of the scene and the 3D model of the virtual character into the virtual reality development platform, assemble the model, and obtain a virtual scene of police-civilian dialogue.

[0083] Based on the assessment question, establish behavioral logic for a 3D model of a virtual character, and set triggers and interaction points;

[0084] The virtual scenario of police-civilian dialogue and the 3D model of the virtual character with established behavioral logic are deployed to VR devices to complete the construction of the police-civilian dialogue scenario questions.

[0085] S5: The subjects are assessed based on police-civilian dialogue scenarios to obtain their language output.

[0086] S6: Use a second language model to analyze the assessment language output and obtain the assessment report of the test taker.

[0087] A further implementation method involves using a second language model to analyze and assess language output as follows:

[0088] Based on the assessment language output, the reaction time of the subjects is collected, and the reaction characteristics of the subjects are extracted. In this embodiment, timing software is used to record the entire process from the subject receiving the test question to making a response, and the reaction time series is obtained. The reaction time series is cleaned, and the subject's reaction time is timestamped with its corresponding assessment language output.

[0089] Then, response features are extracted. Basic statistics such as average reaction time, shortest reaction time, and longest reaction time for each subject are calculated to obtain the first response feature. The distribution of subjects' reaction speeds is analyzed, such as whether it follows a normal distribution or other distribution patterns, to obtain the second response feature. The reaction time of subjects when making mistakes differs from their reaction time when giving correct answers, to obtain the third response feature. The relationship between the complexity of the assessment language output (such as sentence length, vocabulary difficulty, and contextual comprehension requirements) and subjects' reaction times is studied to obtain the fourth response feature. A time-series analysis of these four response features is then performed to obtain the trend of change in subjects' reaction times across different assessment questions.

[0090] The second BERT language model is used to perform syntactic structure analysis and keyword extraction on the test language output, thereby extracting the language features of the test subjects; (the second BERT language model uses the same network architecture as the first BERT language model).

[0091] By combining response characteristics and language characteristics, a comprehensive assessment of the subjects' abilities is conducted. In this embodiment, the subjects' response characteristics and language characteristics are used to evaluate their cognitive load when processing language outputs with different affective tendencies.

[0092] Based on the comprehensive quality assessment results, the subjects were classified into levels and their deficiencies in police work ability were analyzed. In this embodiment, the differences in reaction time and reaction characteristics among different subjects were also explored, and the underlying reasons (such as age, gender, language ability, etc.) were analyzed.

[0093] Based on the grading system and the deficiencies in police work capabilities, an assessment report was obtained for the participants. This report was then provided to the participants or relevant personnel so they could understand their strengths and weaknesses and implement targeted training or improvement accordingly.

[0094] Example 2

[0095] This invention also provides an assessment system and method that combines language models and virtual technology, characterized in that it includes: a practice module, a practice scoring module, an assessment authorization module, an assessment module, and an assessment analysis module;

[0096] The exercise module is used to practice the dialogue process between the police and the public, and the first language model is used to analyze the language output of the subjects in the exercise to obtain the exercise analysis results.

[0097] The exercise scoring module is used to obtain the exercise scores of the subjects based on the exercise analysis results;

[0098] The assessment permission module is used to grant assessment permission when the subject's practice score meets the preset score requirements;

[0099] The assessment module is used to construct police-civilian dialogue scenario questions based on assessment permissions and combined with virtual technology; the subjects are assessed based on the police-civilian dialogue scenario questions to obtain assessment language output;

[0100] The assessment analysis module is used to analyze the assessment language output using a second language model to obtain the assessment evaluation report of the test taker.

[0101] A further implementation method is that the training module includes a training language output acquisition unit, a syntactic information extraction unit, a syntactic information analysis unit, a loss function calculation unit, and a model optimization unit;

[0102] The language output acquisition unit is used to classify and collect problems in daily police work and establish a problem database; it enables subjects to conduct police-civilian dialogue drills based on the problem database and obtain the subjects' drill language output.

[0103] The syntactic information extraction unit is used to extract the syntactic structure information of the practice language output using a tree-structured long short-term memory network and obtain syntactic structure information labels.

[0104] The syntactic information analysis unit is used to input syntactic structure information into the first BERT language model to obtain sentiment classification probability values ​​and keyword information.

[0105] The loss function calculation unit is used to obtain the loss function of syntactic structure information based on syntactic structure information labels, sentiment classification probability values ​​and keyword information;

[0106] The model optimization unit is used to calculate the loss value of syntactic structure information using the loss function, and optimize the parameters of the first BERT language model based on the loss value to obtain the optimal first BERT language model; based on the optimal first BERT language model, the exercise analysis results are obtained.

[0107] A further implementation method is that the assessment module includes: a question extraction unit, a 3D model creation unit, a model assembly unit, a behavioral logic design unit, and a deployment unit;

[0108] The question extraction unit is used to extract questions from the assessment question bank to obtain assessment questions;

[0109] The 3D model creation unit is used to create 3D scene models and 3D virtual character models that meet the assessment questions.

[0110] The model assembly unit is used to import the 3D model of the scene and the 3D model of the virtual character into the virtual reality development platform, assemble the model, and obtain a virtual scene of police-civilian dialogue.

[0111] The behavioral logic design unit is used to establish behavioral logic for the 3D model of a virtual character based on the assessment questions, and to set triggers and interaction points.

[0112] The deployment unit is used to deploy the virtual scene of police-civilian dialogue and the 3D model of the virtual character after establishing the behavioral logic to the VR device, thus completing the construction of the police-civilian dialogue scene questions.

[0113] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made to the technical solutions of the present invention by those skilled in the art without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. An assessment method combining language models and virtual technology, characterized in that, Includes the following steps: The drill involved a dialogue between police and civilians, and the first language model was used to analyze the language output of the participants to obtain the analysis results. Based on the analysis results of the exercise, the exercise score of the subjects was obtained; When the subject's practice score meets the preset score requirement, the subject is granted permission to perform the assessment. Based on the aforementioned assessment and authorization, and combined with virtual technology, a scenario-based question for police-citizen dialogue is constructed. The subjects were assessed based on the police-civilian dialogue scenario questions, and their assessment language output was obtained. The assessment language output is analyzed using a second language model to obtain the subject's assessment evaluation report; The method for obtaining the exercise analysis results is as follows: Collect problems from daily police work by category and establish a problem database; The subjects were asked to conduct a police-civilian dialogue exercise based on the aforementioned question database, and the subjects' exercise language output was obtained. Using a tree-structured long short-term memory network, the syntactic structure information of the exercise language output is extracted, and the syntactic structure information labels are obtained; The syntactic structure information is input into the first BERT language model to obtain the sentiment classification probability value and keyword information; Based on the syntactic structure information labels, the sentiment classification probability values, and the keyword information, a loss function for the syntactic structure information is obtained. The loss value of the syntactic structure information is calculated using a loss function, and the parameters of the first BERT language model are optimized based on the loss value to obtain the optimal first BERT language model; Based on the optimal first BERT language model, the exercise analysis results are obtained; The second BERT language model uses the same network architecture as the first BERT language model.

2. The assessment method combining language models and virtual technology according to claim 1, characterized in that, The results of the exercise analysis include the analysis results of the subjects' ability to recognize intent, the analysis results of communication clarity, the analysis results of communication logical coherence, and the analysis results of the accuracy of application of regulations.

3. The assessment method combining language models and virtual technology according to claim 1, characterized in that, The method for constructing police-citizen dialogue scenario questions is as follows: Test questions are extracted from the test question bank to obtain the test questions; Create a 3D model of the scene and a 3D model of the virtual character that match the assessment questions; The 3D model of the scene and the 3D model of the virtual character are imported into the virtual reality development platform for model assembly to obtain a virtual scene of police-civilian dialogue. Based on the assessment questions, establish behavioral logic for the 3D model of the virtual character, and set triggers and interaction points; The virtual scenario of police-civilian dialogue and the 3D model of the virtual character after establishing behavioral logic are deployed to a VR device to complete the construction of the police-civilian dialogue scenario.

4. The assessment method combining language models and virtual technology according to claim 3, characterized in that, The method for analyzing the assessment language output using a second language model is as follows: Based on the output of the assessment language, the reaction time of the subjects is collected, and the reaction characteristics of the subjects are extracted. Based on the second BERT language model, syntactic structure analysis and keyword extraction are performed on the assessment language output to extract the language features of the subjects. The subjects were assessed on their overall competence by combining the aforementioned response characteristics and language characteristics. Based on the comprehensive quality assessment results, the subjects were classified into different levels and their deficiencies in police work ability were analyzed. Based on the aforementioned grading and the deficiencies in police work capabilities, an assessment report for the subjects was obtained.

5. An assessment system combining language models and virtual technology, employing the method described in any one of claims 1-4, characterized in that, include: The module includes a drill module, a drill scoring module, an assessment and approval module, an assessment module, and an assessment analysis module. The training module is used to practice police-civilian dialogue and uses a first language model to analyze the language output of the subjects during the training to obtain training analysis results. The exercise scoring module is used to obtain the subject's exercise score based on the exercise analysis results; The assessment permission module is used to obtain assessment permission when the subject's practice score meets the preset score requirements; The assessment module is used to construct police-citizen dialogue scenario questions based on the assessment permission and in conjunction with virtual technology. The subjects were assessed based on the police-civilian dialogue scenario questions, and their assessment language output was obtained. The assessment analysis module is used to analyze the assessment language output using a second language model to obtain the assessment evaluation report of the subject.

6. The assessment system combining language models and virtual technology according to claim 5, characterized in that, The training module includes a training language output acquisition unit, a syntactic information extraction unit, a syntactic information analysis unit, a loss function calculation unit, and a model optimization unit; The language output acquisition unit is used to classify and collect daily police work issues and establish a problem database; it enables subjects to conduct police-civilian dialogue drills based on the problem database and obtains the subjects' drill language output. The syntactic information extraction unit is used to extract the syntactic structure information of the exercise language output using a tree-structured long short-term memory network, and to obtain the syntactic structure information label. The syntactic information analysis unit is used to input the syntactic structure information into the first BERT language model to obtain the sentiment classification probability value and keyword information; The loss function calculation unit is used to obtain the loss function of the syntactic structure information based on the syntactic structure information label, the sentiment classification probability value, and the keyword information; The model optimization unit is used to calculate the loss value of the syntactic structure information using a loss function, and optimize the parameters of the first BERT language model based on the loss value to obtain the optimal first BERT language model; and obtain the exercise analysis results based on the optimal first BERT language model.

7. The assessment system combining language models and virtual technology according to claim 5, characterized in that, The assessment module includes: a question extraction unit, a 3D model creation unit, a model assembly unit, a behavioral logic design unit, and a deployment unit; The question extraction unit is used to extract questions from the assessment question bank to obtain assessment questions. The 3D model creation unit is used to create a scene 3D model and a virtual character 3D model that conform to the assessment questions. The model assembly unit is used to import the scene 3D model and the virtual character 3D model into the virtual reality development platform to assemble the model and obtain a virtual scene of police-civilian dialogue. The behavioral logic design unit is used to establish behavioral logic for the three-dimensional model of the virtual character based on the assessment questions, and to set triggers and interaction points. The deployment unit is used to deploy the virtual scene of police-civilian dialogue and the 3D model of the virtual character after establishing behavioral logic to the VR device, thereby completing the construction of the police-civilian dialogue scene questions.

Images