Emergency exercise system and method for emergencies

Abstract

The invention discloses an emergency exercise system and method, and the method comprises the steps: extracting an exercise process record from an evaluation module, carrying out the segmentation processing and exercise evaluation of the exercise process record through employing a data text analysis and coding technology based on artificial intelligence and a large language model, and then, carrying out the segmentation processing and exercise evaluation of the exercise process record; and performing semantic coding on each drill record segment content evaluation text obtained by evaluation, so as to intelligently generate a drill evaluation report according to the jump degree context semantic association representation between the semantics of each drill record segment content evaluation text. According to the method, the evaluation standard is dynamically adjusted by using the AI, and specific requirements can be adaptively identified and responded according to different drilling situations. The flexibility ensures that the evaluation process is closer to an actual drilling scene, and the comprehensiveness and accuracy of evaluation are improved. And meanwhile, the subjective difference caused by human factors is reduced through the intelligently generated evaluation report, and the consistency of evaluation results is ensured.

Description

Technical Field

[0001] The present application relates to the field of intelligent assessment, and more specifically, to an emergency response drill system and method for emergencies. Background Art

[0002] In today's society, various emergencies such as natural disasters, public health incidents, and safety accidents occur frequently, posing serious threats to people's lives, property, and social stability. In order to effectively improve the ability to respond to emergencies, emergency drills have become an indispensable means.

[0003] Patent CN104143130A proposes an emergency response drill system and drill method for emergencies, which includes multiple dedicated terminals and servers for simulating emergencies. The administrator terminal manages authority and maintains the system, the script editor terminal is responsible for material and information management, the director terminal controls the drill progress and role allocation, the student terminal provides drill information and supports interaction, the simulator terminal dynamically adjusts events, and the assessor terminal records and evaluates the drill to obtain a drill evaluation report.

[0004] In this patent, the drill evaluation report is initially evaluated through automation, and finally summarized and written by the evaluator through manual analysis and judgment. However, traditional automated evaluation relies on preset static standards, which is difficult to adapt to a variety of emergencies and drill scenarios, and cannot be flexibly adjusted to meet the emergency response needs in different scenarios. Such fixed standards may ignore the unique requirements of specific situations, resulting in incomplete and inaccurate evaluations. In addition, although the automated evaluation provides preliminary results, the manual review of the evaluator inevitably incorporates personal experience and preferences, which increases the subjectivity and inconsistency of the evaluation results. The evaluator has a large workload and is prone to fatigue, which further affects the quality and efficiency of the evaluation.

[0005] Therefore, an optimized emergency response drill plan for emergencies is desired. Summary of the invention

[0006] In order to solve the above-mentioned technical problems, the present application is proposed. The embodiment of the present application provides an emergency drill system and method for sudden incidents, which extracts the drill process record from the evaluation module, uses data text analysis and coding technology based on artificial intelligence and large language models to perform segmented processing and drill evaluation on the drill process record, and then semantically encodes the content evaluation text of each drill record segment obtained by the evaluation, so as to intelligently generate a drill evaluation report based on the jump degree context semantic association representation between the semantics of the content evaluation text of each drill record segment. The present application uses AI to dynamically adjust the evaluation criteria, and can adaptively identify and respond to specific needs according to different drill scenarios. This flexibility ensures that the evaluation process is closer to the actual drill scenario and improves the comprehensiveness and accuracy of the evaluation. At the same time, the intelligently generated evaluation report reduces the subjective differences caused by human factors and ensures the consistency of the evaluation results.

[0007] According to one aspect of the present application, a system for emergency drills for sudden incidents is provided, which includes: a permission management module for managing user permissions; a script management module for storing and managing various drill script information in emergency drills; a drill control module for monitoring and managing the progress of emergency drills; an information provision module for providing relevant information of emergency drills; a content adjustment module for dynamically adjusting the drill content according to the progress and events of the emergency drills; an evaluation module for recording and analyzing the emergency drill process to obtain a drill evaluation report; and a maintenance module for providing the drill evaluation report to a terminal and performing data maintenance.

[0008] According to another aspect of the present application, a method for emergency drills for sudden incidents is provided, which includes: managing user permissions; storing and managing various drill script information in emergency drills; monitoring and managing the progress of emergency drills; providing relevant information of emergency drills; dynamically adjusting the drill content according to the progress and events of the emergency drills; recording and analyzing the emergency drill process to obtain a drill evaluation report; providing the drill evaluation report to a terminal, and performing data maintenance.

[0009] Compared with the prior art, the present application provides an emergency drill system and method for emergencies, which extracts the drill process record from the evaluation module, uses data text analysis and coding technology based on artificial intelligence and large language models to segment the drill process record and conduct drill evaluation, and then semantically encodes the content evaluation text of each drill record segment obtained by evaluation, so as to intelligently generate a drill evaluation report based on the jump degree context semantic association representation between the semantics of the content evaluation text of each drill record segment. The present application uses AI to dynamically adjust the evaluation criteria, and can adaptively identify and respond to specific needs according to different drill scenarios. This flexibility ensures that the evaluation process is closer to the actual drill scenario and improves the comprehensiveness and accuracy of the evaluation. At the same time, the intelligently generated evaluation report reduces the subjective differences caused by human factors and ensures the consistency of the evaluation results. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0011] Figure 1 4 is a block diagram of an emergency response drill system for emergencies according to an embodiment of the present application.

[0012] Figure 2 Schematic diagram of data flow of an emergency response drill system according to an embodiment of the present application.

[0013] Figure 3 4 is a block diagram of an evaluation module in an emergency response drill system according to an embodiment of the present application.

[0014] Figure 4 The present invention is a flowchart of an emergency response drill method for unexpected events according to an embodiment of the present application. DETAILED DESCRIPTION

[0015] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0016] As shown in this application and claims, unless the context clearly indicates an exception, the words "a", "an", "an" and / or "the" do not refer to the singular and may also include the plural. Generally speaking, the terms "include" and "comprise" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0017] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules can be used and run on the user terminal and / or server. The modules are only illustrative, and different aspects of the system and method can use different modules.

[0018] Flowcharts are used in the present application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps may be processed in reverse order or simultaneously as required. Meanwhile, other operations may also be added to these processes, or a certain step or several steps of operations may be removed from these processes.

[0019] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0020] In the technical solution of the present application, an emergency response drill system for emergencies is proposed. Figure 1 4 is a block diagram of an emergency response drill system for emergencies according to an embodiment of the present application. Figure 2 Schematic diagram of data flow of the emergency response drill system according to an embodiment of the present application. Figure 1 and Figure 2 As shown, according to the embodiment of the present application, the emergency response drill system 300 includes: a permission management module 310 for managing user permissions; a script management module 320 for storing and managing various drill script information in emergency drills; a drill control module 330 for monitoring and managing the progress of emergency drills; an information provision module 340 for providing relevant information of emergency drills; a content adjustment module 350 for dynamically adjusting the drill content according to the progress and events of the emergency drill; an evaluation module 360 ​​for recording and analyzing the emergency drill process to obtain a drill evaluation report; and a maintenance module 370 for providing the terminal with the drill evaluation report and performing data maintenance.

[0021] In particular, the authority management module 310 is used to manage user authority. The operation of the system begins with strict management of user authority, which is a basic step to ensure the security and confidentiality of the drill. Only authorized personnel can access specific functional modules and information resources, which not only protects sensitive data, but also prevents non-related personnel from interfering with the drill process.

[0022] In one example, first, when a user accesses the system for the first time, the registration process will be completed according to the organization to which he belongs and his personal identity information. At this time, the system administrator will assign one or more roles to the new user according to the pre-set role assignment rules, such as script editor, facilitator, trainee, simulator or assessor. Each role is given certain operating permissions, which define what operations the user can perform in the system, such as creating and editing exercise scripts, controlling exercise progress, participating in exercise interactions, or recording and analyzing exercise results. As users use it daily, the system will continue to verify the user's login status and permission level to ensure that each access is legal and controlled. When a user tries to perform a specific function, the system background will automatically check the user's permission configuration to determine whether the user is allowed to perform the operation. If the permission is insufficient, the system will politely reject the request and give a prompt; otherwise, it will be released smoothly. In order to adapt to changes in the internal structure of the organization or adjustments to personal responsibilities, the permission management system also provides a flexible change mechanism. This allows administrators to update the user's permission settings in a timely manner according to actual needs, such as adding or removing certain operating permissions, or changing the user's role. Such flexibility not only helps maintain the security of the system, but also ensures that all participants can obtain the necessary resources and support to complete their tasks in a timely manner. In addition, permission management also includes an audit tracking function. The system will record in detail the time point of each permission change, the objects involved, and the specific reasons for the change, forming a complete log. This is very helpful for post-event review and responsibility tracing, and can effectively prevent the abuse of power.

[0023] In particular, the script management module 320 is used to store and manage various drill script information in emergency drills. These scripts are the core of the drills, and the drill script information contains various possible event scenarios, interaction problems, and detailed project designs. In this way, the system can provide a rich variety of drill scenarios, allowing participants to practice decision-making and response skills in a near-real environment. At the same time, this also provides flexibility for drills, and the script content can be customized according to different training needs.

[0024] In one example, in an emergency drill system, users first need to be provided with an easy-to-use interface to write or import drill scripts. This interface should be intuitive and feature-rich, allowing script editors to design detailed drill processes based on actual emergency types (such as natural disasters, public health events, safety accidents, etc.). These scripts are not just simple text descriptions, but also include drill goal setting, role allocation, timeline planning, expected action steps, decision points, and possible scenario change branches. Once the script is written and uploaded to the system, the next step is to properly store and organize it. A dedicated database is set up inside the system to save all versions of drill scripts, and metadata is used to mark the key attributes of each script, such as type, scope of application, creation date, and author information. This structured storage method not only facilitates subsequent retrieval queries, but also provides a basis for version control, ensuring that each update iteration can be accurately recorded to facilitate tracking of historical changes.

[0025] In particular, the drill control module 330 is used to monitor and manage the progress of the emergency drill. Throughout the process, the system integrates a variety of technical means and tools to provide a powerful platform for the instructor, enabling him to fully control the dynamics of the drill, adjust strategies in a timely manner, and ensure the authenticity and effectiveness of the drill.

[0026] In one example, before the drill began, the system had already developed a detailed drill plan based on the preset script and the role allocation of the participants. This plan covers key elements such as the time arrangement of the drill, the selection of the location, and the definition of the roles of the participating units and personnel. In order to ensure that all participants can clearly understand their tasks, the system will send notifications in advance and open online training resources to help everyone familiarize themselves with the process and rules. At the same time, the system also sets up a simulation test environment, allowing users to rehearse before the formal drill to verify the adequacy of the preparations. When the drill is officially launched, the system's monitoring function is immediately activated. The instructor can observe the progress of the drill in real time through a dedicated terminal, and the screen displays status feedback from each participating terminal, including geographic location information, communication records, task completion status, etc. These data are not only presented in the form of intuitive charts, but can also be visualized through GIS maps, making the entire drill scene clear at a glance. In addition, the system has an intelligent early warning mechanism. Once a deviation from the predetermined path or abnormal behavior is detected, it will immediately issue an alarm to the instructor to remind him of potential problems. The system provides a high degree of flexibility for changes that may occur during the drill. If an emergency causes the original script to be unable to continue, the director can make decisions quickly according to the actual situation, modify the drill content or re-plan the course of action. This ability to make instant adjustments is due to the system's built-in powerful editor and a database that supports multi-version control, so that any changes can be quickly deployed without affecting the continuity of the overall drill. In addition to technical support, communication and collaboration between people are also crucial. To this end, the system has built an efficient communication network that connects all participating terminals to ensure unimpeded information flow. Whether it is the transmission of instructions between the command center and the on-site commander, or the coordination and cooperation between different groups, seamless docking can be achieved on this platform. As the drill progresses, the system continues to collect various indicator data, such as response time, resource allocation efficiency, public response, etc. These data are not only the basis for evaluating the effectiveness of the drill, but also provide a basis for subsequent improvements.

[0027] In particular, the information providing module 340 is used to provide relevant information about the emergency drill. In this process, in order to support the effective participation of trainees, the system needs to provide detailed information related to the emergency drill, including but not limited to multimedia display and real-time communication functions. This information helps trainees better understand the current situation and make more accurate judgments. At the same time, the simulator can increase the realism and challenge of the drill by temporarily inserting or deleting events, further testing the reaction speed and collaboration ability of the participants.

[0028] In one example, during the preparation stage of an exercise, the system will release notifications of upcoming exercises to all potential participants through multiple channels. This includes not only formal documents and emails, but also social media, instant messaging tools, and dedicated applications. These notifications detail the purpose, schedule, location selection, and basic rules of the exercise, helping participants to prepare mentally and technically in advance. In addition, the system also provides online training modules, including resources such as video tutorials, documentation, and simulation exercises, so that users can self-study at any time and place to familiarize themselves with the exercise process and their respective tasks. As the exercise date approaches, the system further refines the content of the information release, especially specific guidance for different roles. For example, for script editors, the system will provide professional advice on how to create complex scenario branches and decision points; for facilitators, the focus is on exercise progress control techniques and emergency handling methods; and for trainees and evaluators, the system emphasizes the ways to participate in interaction and evaluation criteria. This personalized information service ensures that each participant can get the most suitable help for themselves and improve the actual effect of the exercise. After entering the exercise implementation stage, the system's real-time information update function becomes particularly important. It is not only necessary to be able to display the current status of the exercise, such as ongoing tasks, the location distribution of each group, etc., but also to dynamically adjust the display content to adapt to the changing situation. In order to ensure the efficiency of information transmission, the system adopts a multi-layer filtering mechanism to filter out the most relevant information and push it to users according to their authority level and focus. For example, front-line operators mainly receive operating instructions directly related to themselves, while management can obtain a broader global perspective to make strategic decisions. During the entire exercise, the system also set up a special information exchange platform to encourage interactive communication between participants. Especially when encountering emergencies, fast and effective information exchange can greatly improve the speed and quality of problem solving. After the exercise, the system automatically generates a detailed summary report and distributes it to each participant. This report not only covers the overall review of the exercise, but also includes performance evaluation of each key node and recognition of individual contributions.

[0029] In particular, the content adjustment module 350 is used to dynamically adjust the drill content according to the progress and events of the emergency drill. For example, the drill director uses a dedicated terminal to control the progress of the drill, assigns roles according to actual conditions, and sends event messages or questions to participants in a timely manner. In this process, the system must also be able to dynamically adjust the drill content to reflect the unpredictability of emergencies. For example, when simulating a rescue operation after an earthquake, if a new building collapses or a fire occurs, the system should be able to quickly insert related events, allowing participants to adjust their strategies immediately, thereby improving their on-the-spot response capabilities.

[0030] In one example, as the drill progresses, if it is found that the actual situation deviates from the predetermined script, or an unforeseen new situation arises, the system will immediately start the dynamic adjustment mechanism. At this time, the instructor can use the system's real-time editing function to respond quickly and adjust the drill content. For example, if the weather in a certain area suddenly deteriorates and affects the rescue operation, the instructor can quickly change the scenario setting of the relevant area through the system, introduce new challenges, and require participants to reassess the situation and take corresponding measures. This ability to make instant adjustments is due to the system's built-in powerful editor and database that supports multi-version control, so that any changes can be deployed quickly without affecting the continuity of the overall drill.

[0031] In particular, the evaluation module 360 ​​is used to record and analyze the emergency drill process to obtain a drill evaluation report. Figure 3 As shown, the evaluation module 360 ​​includes: a rehearsal process record extraction unit 361, which is used to extract the rehearsal process record from the evaluation module; a record segmentation evaluation unit 362, which is used to segment the rehearsal process record and perform rehearsal evaluation to obtain the sequence distribution of the rehearsal record segment content evaluation text; a record encoding unit 363, which is used to semantically encode the sequence distribution of the rehearsal record segment content evaluation text to obtain the sequence distribution of the rehearsal record segment content evaluation text semantic encoding vector; a record context semantic association unit 364, which is used to perform context semantic association of feature jump degree on the sequence distribution of the rehearsal record segment content evaluation text semantic encoding vector to obtain the context semantic association encoding vector between the rehearsal record segment content evaluation text; and a rehearsal evaluation report generation unit 365, which is used to obtain the rehearsal evaluation report based on the context semantic association encoding vector between the rehearsal record segment content evaluation text.

[0032] Specifically, the drill process record extraction unit 361 is used to extract the drill process record from the evaluation module. Among them, the drill process record contains rich multi-dimensional information, from geographic location to time series, from physical behavior to communication, and almost covers all activities that occur during the drill. This comprehensive feature allows evaluators to examine the drill effect from multiple angles and find problems that are difficult to detect from a single perspective. It is worth mentioning that the drill process record comes from the real-time data stream provided by each participating terminal. Before the drill begins, each participating unit and individual will be assigned a specific role and equipped with corresponding communication equipment and technical tools. These devices include but are not limited to GPS positioning devices, radio communicators, mobile applications, etc., which continuously send various forms of data such as location information, communication records, task completion status, etc. to the central server throughout the drill. In addition to the automated data collection of hardware devices, the drill process record also relies on information input by humans. For example, the instructor can manually enter the time, location and specific circumstances of important events through a dedicated terminal; trainees can submit a brief self-evaluation or description of the problems encountered after completing the designated task. These manually input data supplement the situational details that the machine cannot capture, making the record more complete and vivid.

[0033] Specifically, the record segmentation evaluation unit 362 is used to segment the drill process record and perform drill evaluation to obtain the sequence distribution of the drill record segment content evaluation text. That is, in the technical solution of the present application, first, the drill process record is segmented to obtain the sequence distribution of the drill process record segment description. Here, considering that the drill process record usually contains a lot of complex information, covering many links and details from the beginning to the end of the drill, including the existence of multiple key nodes and stages, different stages have different tasks and focuses. Therefore, in order to accurately identify and highlight these key nodes, such as the start of emergency response, completion of resource allocation, key decision-making moments, etc., the present application segments the drill process record to divide the long drill process into smaller and more manageable parts, and obtains the sequence distribution of the drill process record segment description. In this way, it helps to focus on these key moments, that is, each stage and link of the drill can be evaluated in a refined manner. For example, for the evacuation link record segment, the evacuation speed and route rationality can be evaluated; for the rescue link record segment, the effectiveness of rescue measures and teamwork can be evaluated. This detailed evaluation helps to fully understand the effect of the drill and discover the advantages and disadvantages of each link. Then, in order to clearly understand whether each drill process record segment description has achieved the expected goals, such as the execution of various emergency plans by the participants, the degree of improvement in their ability to respond to emergencies, etc. In the technical solution of the present application, after adding prompt words to the end of each drill process record segment description in the sequence distribution of the drill process record segment description, it is input into the drill evaluation module based on the large language model to obtain the sequence distribution of the drill record segment content evaluation text, in particular, the prompt words are: conduct a drill evaluation on the content of the drill process record paragraph. It should be understood that the large language model has a wide range of versatility and can handle various types of language tasks. However, for specific drill evaluation tasks, the model needs to be clearly informed of the specific operations to be performed. Adding the prompt "conduct a drill evaluation on the content of the drill process recording paragraph" is like giving the model clear instructions, letting the model know exactly that it needs to conduct a drill evaluation on the input text paragraph, so as to more comprehensively improve the relevance and accuracy of the generated evaluation text, thereby providing a direct data basis for the subsequent generation of a comprehensive and organized drill evaluation report.

[0034] Specifically, the record encoding unit 363 is used to semantically encode the sequence distribution of the drill record segment content evaluation text to obtain the sequence distribution of the semantic encoding vector of the drill record segment content evaluation text. That is, in the technical solution of the present application, a semantic encoder based on the Bert-bidirectional LSTM hybrid model is used to semantically encode each drill record segment content evaluation text in the sequence distribution of the drill record segment content evaluation text to obtain the sequence distribution of the semantic encoding vector of the drill record segment content evaluation text. Considering that each drill record segment content evaluation text contains semantic information of different content segments, and there are implicit emotions and context-related semantic meanings between these semantic information, based on this, in the technical solution of the present application, each drill record segment content evaluation text in the sequence distribution of the drill record segment content evaluation text is semantically encoded to mine the complex information and relationships contained in the text, and obtain the sequence distribution of the semantic encoding vector of the drill record segment content evaluation text. In particular, in one embodiment of the present application, the semantic encoder based on the Bert-bidirectional LSTM hybrid model is used to semantically encode each drill record segment content evaluation text in the sequence distribution of the drill record segment content evaluation text. It can be understood that Bert is a pre-trained language model based on Transformer, which is mainly composed of the encoding layer of Transformer. The encoding layer is based on the self-attention structure and can effectively capture the global information of the text sequence. For example, when processing long documents, it can quickly form an overall grasp of the semantics of each drill record segment content evaluation text, and can understand the general situation of the entire drill process from a macro level. Although Bert performs well in parallel training and capturing global information, the encoding layer based on the self-attention structure has difficulties in obtaining the order and grammatical relationship between words in the text sequence. The bidirectional LSTM processes the sequence from front to back and from back to front, so as to better obtain the contextual information of each word. It is especially good at capturing the order and grammatical relationship between words. For example, when analyzing the sentences describing the sequence of rescue operations in the drill record text, the bidirectional LSTM can clearly understand the logical relationship expressed by the word order. Specifically, in this hybrid model, Bert is first used to perform preliminary processing on the input text, and its powerful ability to extract global features and semantic representations of the text. Then, the features output by Bert are used as the input of the bidirectional LSTM, and these features are further processed in depth to mine the contextual details and sequential information at the word level to more accurately understand the meaning of the drill record text.

[0035] Specifically, the record context semantic association unit 364 is used to perform context semantic association of feature jump degree on the sequence distribution of the semantic coding vector of the content evaluation text of the drill record segment to obtain the context semantic association coding vector between the content evaluation texts of the drill record segment. Considering that the drill process is a continuous whole, there is an inherent logical connection between the record segments. A simple semantic coding vector is only an independent representation of each record segment, and it is difficult to directly reflect the context association between them. Therefore, in order to capture and refine the association between the semantic information between the content evaluation texts of each record segment, so as to reveal how the evaluation texts of each record segment are connected and influenced with each other, and then fully understand the description of the entire drill process, the present application performs context semantic association of feature jump degree on the sequence distribution of the semantic coding vector of the content evaluation text of the drill record segment to fully reflect the dynamic semantic relationship between the texts, and obtains the context semantic association coding vector between the content evaluation texts of the drill record segment. In particular, when information is transmitted, important nodes can be accurately located by analyzing them. For example, in an emergency drill data processing network, the nodes corresponding to key decision points may show unique characteristic jumps and message transmission space spans. After finding these important nodes, the information flow can be dynamically adjusted to enhance the transmission of key information in order to more accurately evaluate the drill effect.

[0036] Specifically, first, the message transmission significance factor of each drill record segment content evaluation text semantic coding vector in the sequence distribution of the drill record segment content evaluation text semantic coding vector is calculated to obtain the sequence distribution of the drill record segment content evaluation text semantic message transmission significance factor. That is, in an embodiment of the present application, first, the feature jump degree of each drill record segment content evaluation text semantic coding vector in the sequence distribution of the drill record segment content evaluation text semantic coding vector is calculated to obtain the sequence distribution of the drill record segment content evaluation text semantic feature jump degree; in this way, the system can capture the local structural information and the mutation characteristics between nodes in the drill record segment content. For example, in the scenario of a military exercise, if a new tactic or equipment is suddenly introduced in a certain exercise stage, the corresponding drill record segment may show a feature jump degree that is significantly different from other segments. This step not only helps to identify which drill record segments have major changes in the network, but also emphasizes the key points that may have an important impact on subsequent operations. Next, in order to further examine the ability of information to spread throughout the entire drill record network, in the technical solution of the present application, the message transmission space span of each drill record segment content evaluation text semantic coding vector in the sequence distribution of the drill record segment content evaluation text semantic coding vector is calculated to obtain the sequence distribution of the drill record segment content evaluation text semantic message transmission space span; by calculating the message transmission space span of each drill record segment content evaluation text semantic coding vector, it is possible to obtain a broader perspective of connectivity and path length considerations. This means that for drill records, the impact of certain decisions or events may go beyond the directly related paragraphs, but have an effect in a wider range. For example, a successful strategic adjustment may affect multiple subsequent drill links, and even affect the results of the entire drill. Then, based on the feature jump degree and message transmission space span of each drill record segment content evaluation text semantic coding vector in the sequence distribution of the drill record segment content evaluation text semantic coding vector, the message transmission significance factor of each drill record segment content evaluation text semantic coding vector is calculated to obtain the sequence distribution of the drill record segment content evaluation text semantic message transmission significance factor. Here, the nodes that play a more important role in the information dissemination process are determined by comprehensively considering the feature jump degree and the spatial span. In this way, it is ensured that the important nodes in the information dissemination process can be highlighted while retaining the overall structural information, and each node is assigned a score that reflects its importance in the information dissemination process. In a specific example, the message transmission significance factor of each drill record segment content evaluation text semantic coding vector in the sequence distribution of the drill record segment content evaluation text semantic coding vector is calculated using the following message transmission significance factor calculation formula to obtain the sequence distribution of the drill record segment content evaluation text semantic message transmission significance factor; wherein the message transmission significance factor calculation formula is: ;in, is the sequence distribution of the semantic encoding vector of the content evaluation text of the drill record segment, and They are the first, second, and third in the sequence distribution of the semantic encoding vector of the content evaluation text of the drill record segment. and The semantic encoding vector of the content evaluation text of the drill record segment, yes Middle The eigenvalues ​​at the positions, yes The number of eigenvalues ​​in , and They are and The corresponding weighted semantic average feature value, It is the first in the sequence distribution of the semantic feature jump degree of the drill record content evaluation text The content of each exercise record segment evaluates the jumpiness of the text semantic features. express arrive The number of message delivery times, is the sequence distribution of the spatial span of the exercise record segment content evaluation text semantic message transmission The content of the exercise recording segment evaluates the text semantic message transmission space span, and is the modulation parameter, is the sequence distribution of the significant factor of the text semantic message transmission in the exercise record content evaluation The content of the exercise recording segments is used to evaluate the semantic message transmission salience factor of the text.

[0037] Furthermore, based on the sequence distribution of the semantic message transmission significance factor of the content evaluation text of the drill record segment, the sequence distribution of the semantic encoding vector of the content evaluation text of the drill record segment is subjected to weighted transfer context fusion to obtain the contextual semantic association encoding vector between the content evaluation texts of the drill record segment. That is, in an embodiment of the present application, first, the sequence distribution of the semantic message transmission significance factor of the content evaluation text of the drill record segment is gated for transfer significance to obtain the sequence distribution of the semantic message transmission significance weight of the content evaluation text of the drill record segment; that is, in order to more accurately control the information flow, the sequence distribution of the semantic message transmission significance factor of the content evaluation text of the drill record segment is input into the gated transfer unit, and the gating mechanism allows the model to selectively remember or ignore part of the input, thereby ensuring that only the most relevant information is passed. In this scenario, this means that the system can better focus on those record segments that really have a decisive impact on the drill results, while appropriately weakening less important features. Finally, based on the sequence distribution of the significant weights of the semantic message transmission of the drill record segment content evaluation text, the weighted sum of the sequence distribution of the semantic coding vector of the drill record segment content evaluation text is calculated to obtain the contextual semantic association coding vector between the drill record segment content evaluation texts. That is, by integrating information from different drill record segments, the contextual semantic association coding vector between the drill record segment content evaluation texts is obtained, which reflects the complex relationship between the drill record segments and emphasizes those factors that are most critical to the overall drill results. In this way, the system can not only identify important nodes and connections in the drill, but also predict future trends and development directions, providing strong support and guidance for command. In a specific example, the sequence distribution of the semantic coding vector of the drill record segment content evaluation text is subjected to weighted transfer context fusion using the following weighted transfer context fusion formula to obtain the contextual semantic association coding vector between the drill record segment content evaluation texts; wherein the weighted transfer context fusion formula is: :in, is the normalization function, is a masking operation, is a predetermined threshold, is the sequence distribution of the significant weight of the text semantic message transmission in the exercise recording segment content evaluation The content of each exercise recorded segment evaluates the semantic message delivery salience weight of the text, is the number of vectors in the sequence distribution of the semantic encoding vector of the content evaluation text of the drill record segment, is a contextual semantic association encoding vector between the content evaluation texts of the drill record segment.

[0038] Specifically, the drill evaluation report generating unit 365 is used to obtain the drill evaluation report based on the context semantic association coding vector between the content evaluation texts of the drill record segment. That is, in the technical solution of the present application, after adding a prompt word to the tail of the context semantic association coding vector between the content evaluation texts of the drill record segment, it is input into the drill record global evaluation polishing module based on the large language model to obtain the drill evaluation report. It should be understood that the context semantic association coding vector between the content evaluation texts of the drill record segment contains rich semantic information, but it is relatively abstract. The prompt word can be used as a supplementary information to help the model better understand the connection between the content represented by the vector and the task of generating a report, so that the model is more targeted when processing these vectors, thereby more effectively generating content that meets the requirements. The large language model has various functions and can perform a variety of natural language processing tasks. Adding prompt words such as "generate drill evaluation report" can clearly inform the model of the specific task, let it focus on generating a drill evaluation report, and avoid the model from generating outputs that deviate from the task. In this way, the drill evaluation report can provide a comprehensive summary of the entire drill process, covering aspects such as the achievement of drill objectives, performance of participants, and rationality of the process. This provides detailed feedback to drill organizers and participants, helping them understand the strengths and weaknesses of the drill. For example, organizers can find bottlenecks in the drill process based on the report, and participants can know where their own operations or decisions can be improved, thereby improving their ability to respond to emergencies.

[0039] In a preferred example, after adding a prompt word to the tail of the context semantic association encoding vector between the content evaluation texts of the drill record segment, inputting it into the drill record global evaluation polishing module based on the large language model to obtain the drill evaluation report includes: based on the feature value of the context semantic association encoding vector between the content evaluation texts of the drill record segment, Distance and The distance matrix of contextual semantic association between the content evaluation texts of the drill record segment is obtained. and drill record segment content evaluation text context semantic association two distance matrices .

[0040] Calculate the weighted sum of the contextual semantic association distance matrix between the content evaluation texts of the drill record segments and the contextual semantic association distance matrix between the content evaluation texts of the drill record segments to obtain the contextual semantic association joint distance matrix between the content evaluation texts of the drill record segments .

[0041] Determine the eigenvalues ​​of the contextual semantic association joint distance matrix between the content evaluation texts of the drill record segment , and compose the drill record segment content evaluation text context semantic association joint distance eigenvector .

[0042] The context semantic association encoding vector between the content evaluation texts of the drill record segment as a row vector is matrix-multiplied with the context semantic association-distance matrix between the content evaluation texts of the drill record segment to obtain the context semantic association-distance query vector between the content evaluation texts of the drill record segment .

[0043] The context semantic association distance matrix between the drill record segment content evaluation texts is matrix-multiplied with the self-association matrix of the context semantic association encoding vector between the drill record segment content evaluation texts to obtain the context semantic association distance association matrix between the drill record segment content evaluation texts .

[0044] After matrix multiplication of the distance query vector of the contextual semantic association between the content evaluation texts of the drill record segment and the distance association matrix of the contextual semantic association between the content evaluation texts of the drill record segment, further dot multiplication is performed with the contextual semantic association joint distance eigenvector of the content evaluation texts of the drill record segment to obtain the optimized contextual semantic association encoding vector of the content evaluation texts of the drill record segment ; and after adding a prompt word to the tail of the contextual semantic association encoding vector between the optimized content evaluation texts of the rehearsal record segment, input it into the rehearsal record global evaluation polishing module based on the large language model to obtain a rehearsal evaluation report.

[0045] Here, in the case where each semantic encoding vector of the content evaluation text of the rehearsal record segment in the sequence of semantic encoding vectors of the content evaluation text of the rehearsal record segment respectively represents the encoded semantic features of the generated semantic fragment of the corresponding content evaluation text of the rehearsal record segment, when performing contextual semantic association of feature jump degree, insufficient correspondence of feature jump degree association in the global semantic space caused by the difference in local semantic space of the generated encoded semantics will cause sparse sequence message association transmission of contextual semantic association encoding vectors between the content evaluation texts of the rehearsal record segment, thereby reducing the semantic quality of the rehearsal evaluation report obtained by inputting the global evaluation polishing module of the rehearsal record based on the large language model due to the lack of generated reasoning degree.

[0046] Therefore, the first distance matrix and the second distance matrix of the contextual semantic association coding vector between the content evaluation texts of the rehearsal record segment are used as the fine-grained metric association cluster representation of the contextual semantic association coding vector between the content evaluation texts of the rehearsal record segment, and the dynamic programming of the relationship between the association clusters of different association clusters is performed on the contextual semantic association coding vector between the content evaluation texts of the rehearsal record segment and the self-association representation of the contextual semantic association coding vector between the content evaluation texts of the rehearsal record segment respectively to simulate the sparse activation based on neuron clusters of the association system, and the fine-grained predictable sparsity of the contextual semantic association coding vector between the content evaluation texts of the rehearsal record segment is coordinated with the intrinsic representation of the metric association cluster of the first distance matrix and the second distance matrix of the contextual semantic association coding vector between the content evaluation texts of the rehearsal record segment is performed, so as to avoid the lack of association caused by sparsity affecting the generation of reasoning degree loss, and improve the semantic quality of the rehearsal evaluation report obtained by inputting the contextual semantic association coding vector between the content evaluation texts of the rehearsal record into the rehearsal record global evaluation polishing module based on the large language model.

[0047] In particular, the maintenance module 370 is used to provide the terminal with the exercise evaluation report and perform data maintenance. In this way, it not only provides valuable data support for future exercise improvements, but also provides feedback for participating units and individuals. By continuously optimizing the exercise plan, the system is committed to continuously improving the level of emergency response, ensuring that people can respond to challenges more calmly in the face of real emergencies, and maximizing the protection of life and property safety and social stability. In terms of data maintenance, the system has taken a series of measures to ensure the accuracy and completeness of the exercise data. First, all raw data will be automatically backed up to cloud storage to prevent data loss due to hardware failure or other unexpected situations. Secondly, the system regularly performs data cleaning tasks to remove possible noise or outliers and maintain the purity of the data set. In addition, in order to adapt to the ever-changing exercise needs and technological advances, the system also has a flexible data structure update capability, allowing administrators to adjust field definitions and format requirements according to actual conditions to ensure that the data model always meets the latest standards.

[0048] As described above, the emergency response drill system 300 according to the embodiment of the present application can be implemented in various wireless terminals, such as a server with an emergency response drill algorithm. In a possible implementation, the emergency response drill system 300 according to the embodiment of the present application can be integrated into the wireless terminal as a software module and / or a hardware module. For example, the emergency response drill system 300 can be a software module in the operating system of the wireless terminal, or can be an application developed for the wireless terminal; of course, the emergency response drill system 300 can also be one of the many hardware modules of the wireless terminal.

[0049] Alternatively, in another example, the emergency response drill system 300 and the wireless terminal may be separate devices, and the emergency response drill system 300 may be connected to the wireless terminal via a wired and / or wireless network and transmit interactive information in a predetermined data format.

[0050] Furthermore, a method for emergency response drill for emergencies is also provided.

[0051] Figure 4 Flow chart of the emergency response drill method for emergencies according to an embodiment of the present application. Figure 4 As shown, the emergency drill method for sudden incidents according to the embodiment of the present application includes the following steps: S1, managing user permissions; S2, storing and managing various drill script information in emergency drills; S3, monitoring and managing the progress of emergency drills; S4, providing relevant information of emergency drills; S5, dynamically adjusting the drill content according to the progress and events of the emergency drills; S6, recording and analyzing the emergency drill process to obtain a drill evaluation report; S7, providing the drill evaluation report to the terminal and performing data maintenance.

[0052] In summary, the emergency response drill method for emergencies according to the embodiment of the present application is explained, which extracts the drill process record from the evaluation module, uses data text analysis and coding technology based on artificial intelligence and large language models to perform segmented processing and drill evaluation on the drill process record, and then semantically encodes the content evaluation text of each drill record segment obtained by evaluation, so as to intelligently generate a drill evaluation report based on the jump degree context semantic association representation between the semantics of the content evaluation text of each drill record segment. The present application uses AI to dynamically adjust the evaluation criteria, which can adaptively identify and respond to specific needs according to different drill scenarios. This flexibility ensures that the evaluation process is closer to the actual drill scenario and improves the comprehensiveness and accuracy of the evaluation. At the same time, the intelligently generated evaluation report reduces the subjective differences caused by human factors and ensures the consistency of the evaluation results.

[0053] The embodiments of the present disclosure have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of terms used herein is intended to best explain the principles of the embodiments, practical applications, or improvements to the technology in the market, or to enable other persons of ordinary skill in the art to understand the embodiments disclosed herein.

Claims

1. An emergency drill system for emergencies, comprising: The permission management module is used to manage user permissions; The script management module is used to store and manage various drill script information in emergency drills; Exercise control module, used to monitor and manage the progress of emergency exercises; An information provision module, used to provide relevant information of the emergency drill; a content adjustment module, used to dynamically adjust the drill content according to the progress and events of the emergency drill; Evaluation module, used to record and analyze the emergency drill process to obtain a drill evaluation report; A maintenance module, used to provide the terminal with the rehearsal evaluation report and perform data maintenance; characterized in that the evaluation module includes: a rehearsal process record extraction unit, used to extract the rehearsal process record from the evaluation module; a record segmentation evaluation unit, used to perform segmentation processing and rehearsal evaluation on the rehearsal process record to obtain the sequence distribution of the rehearsal record segment content evaluation text; a record encoding unit, used to semantically encode the sequence distribution of the rehearsal record segment content evaluation text to obtain the sequence distribution of the rehearsal record segment content evaluation text semantic encoding vector; a record context semantic association unit, used to perform context semantic association of feature jump degree on the sequence distribution of the rehearsal record segment content evaluation text semantic encoding vector to obtain the context semantic association encoding vector between the rehearsal record segment content evaluation text; a rehearsal evaluation report generation unit, used to obtain the rehearsal evaluation report based on the context semantic association encoding vector between the rehearsal record segment content evaluation text.

2. The emergency response drill system according to claim 1, characterized in that: The record segmentation evaluation unit includes: a segmentation subunit, which is used to segment the rehearsal process record to obtain a sequence distribution of the rehearsal process record segment description; a rehearsal record segment content evaluation subunit, which is used to add prompt words to the end of each rehearsal process record segment description in the sequence distribution of the rehearsal process record segment description, and then input it into a rehearsal evaluation module based on a large language model to obtain a sequence distribution of the rehearsal record segment content evaluation text.

3. The emergency response drill system according to claim 2, characterized in that: The prompt words are: conduct a drill evaluation on the content of the drill process recording paragraph.

4. The emergency response drill system according to claim 3, characterized in that: The record encoding unit is used to: use a semantic encoder based on a Bert-bidirectional LSTM hybrid model to semantically encode each exercise record segment content evaluation text in the sequence distribution of the exercise record segment content evaluation text to obtain a sequence distribution of the semantic encoding vectors of the exercise record segment content evaluation text.

5. The emergency response drill system according to claim 4, characterized in that: The record context semantic association unit includes: a rehearsal content message transmission significance factor calculation subunit, which is used to calculate the message transmission significance factor of each rehearsal record segment content evaluation text semantic coding vector in the sequence distribution of the rehearsal record segment content evaluation text semantic coding vector to obtain the sequence distribution of the rehearsal record segment content evaluation text semantic message transmission significance factor; a rehearsal record transmission fusion subunit, which is used to perform weighted transmission context fusion on the sequence distribution of the rehearsal record segment content evaluation text semantic coding vector based on the sequence distribution of the rehearsal record segment content evaluation text semantic message transmission significance factor to obtain the context semantic association coding vector between the rehearsal record segment content evaluation texts.

6. The emergency response drill system according to claim 5, characterized in that: The rehearsal content message transmission significance factor calculation subunit is used to: calculate the feature jump degree of each rehearsal record segment content evaluation text semantic coding vector in the sequence distribution of the rehearsal record segment content evaluation text semantic coding vector to obtain the sequence distribution of the rehearsal record segment content evaluation text semantic feature jump degree; calculate the message transmission space span of each rehearsal record segment content evaluation text semantic coding vector in the sequence distribution of the rehearsal record segment content evaluation text semantic coding vector to obtain the sequence distribution of the rehearsal record segment content evaluation text semantic message transmission space span; based on the feature jump degree and message transmission space span of each rehearsal record segment content evaluation text semantic coding vector in the sequence distribution of the rehearsal record segment content evaluation text semantic coding vector, calculate the message transmission significance factor of each rehearsal record segment content evaluation text semantic coding vector to obtain the sequence distribution of the rehearsal record segment content evaluation text semantic message transmission significance factor.

7. The emergency response drill system according to claim 6, characterized in that: The rehearsal record transfer fusion sub-unit is used to: perform gated transfer saliency on the sequence distribution of the semantic message transfer significance factors of the rehearsal record segment content evaluation text to obtain the sequence distribution of the semantic message transfer significance weights of the rehearsal record segment content evaluation text; based on the sequence distribution of the semantic message transfer significance weights of the rehearsal record segment content evaluation text, calculate the weighted sum of the sequence distribution of the semantic encoding vector of the rehearsal record segment content evaluation text to obtain the contextual semantic association encoding vector between the rehearsal record segment content evaluation texts.

8. The emergency response drill system according to claim 7, characterized in that: The rehearsal evaluation report generating unit is used to: after adding a prompt word to the tail of the context semantic association encoding vector between the content evaluation texts of the rehearsal record segment, input it into the rehearsal record global evaluation polishing module based on the large language model to obtain the rehearsal evaluation report.

9. A method for emergency drills for emergencies, characterized in that: include: Manage user permissions; store and manage various drill script information in emergency drills; Monitor and manage the progress of emergency drills; Provide relevant information of emergency drills; dynamically adjust the drill content according to the progress and events of the emergency drills; record and analyze the emergency drill process to obtain a drill evaluation report; provide the terminal with the drill evaluation report and perform data maintenance.

Images