AI-based intelligent crime investigation teaching system and method

Abstract

The invention relates to the technical field of teaching systems, and discloses an AI-based intelligent crime investigation teaching system and method, and the system comprises a data management module which is used for collecting, processing and managing crime investigation related data, and guaranteeing the diversity and quality of the data; the case generation and management module is used for dynamically generating and managing teaching cases based on artificial intelligence, and the teaching cases comprise case backgrounds, clue relationships and reasoning paths; the intelligent learning and reasoning module is used for supporting students to learn crime investigation knowledge and carry out case reasoning and practical operation; and the user interaction module is used for providing a multi-mode interaction mode and supporting a personalized learning path and dynamic feedback. By recording the reasoning path of the student in real time, generating a case draft and dynamically updating the case draft to a case database after auditing and optimizing by a lecturer, a continuous extension mechanism of teaching resources is formed, the innovative thinking and participation enthusiasm of the student are stimulated, and the interactivity of the system and the practical application value of learning results are enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of teaching systems, and specifically to an AI-based intelligent crime investigation teaching system and method. Background Art

[0002] The AI-based intelligent crime investigation teaching system and method is a device designed by combining artificial intelligence technology and the needs of crime investigation education. It is mainly used to assist students in learning investigative reasoning and cultivating case analysis capabilities in a virtual environment. The device is widely used in the fields of public security, legal education, and investigative technology training, providing an intelligent teaching platform for improving the practical capabilities of investigators.

[0003] In the existing technology, there are limited channels for collecting teaching cases, which mainly rely on manual design or acquisition from fixed case libraries. There is a lack of real cases generated by students in practice as a resource supplement, which makes it difficult to dynamically expand the case library. Due to the lack of means to collect teaching cases, the existing system relies on fixed templates and a single structure when dynamically combing case clues and providing voice prompts, and cannot adapt to the real-time requirements of clue diversity and logical associations in complex cases. This limitation makes it difficult for the in-depth guidance module to provide accurate support during the student's reasoning process, weakening the system's effectiveness in enhancing the immersive experience and improving case analysis capabilities. Summary of the invention

[0004] In view of the shortcomings of the existing technology, the present invention provides an AI-based intelligent crime investigation teaching system and method, which solves the problems of lack of channels for the system to obtain case teaching and difficulty in providing accurate support in guiding students' reasoning process.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: an AI-based intelligent crime investigation teaching system, comprising:

[0006] Data management module, used to collect, process and manage crime investigation related data to ensure data diversity and quality;

[0007] Case generation and management module, which is used to dynamically generate and manage teaching cases based on artificial intelligence, including case background, clue relationships and reasoning paths;

[0008] Intelligent learning and reasoning module, used to support students in learning criminal investigation knowledge, carrying out case reasoning and practical operations;

[0009] User interaction module, which is used to provide multimodal interaction and support personalized learning paths and dynamic feedback;

[0010] The practical information collection module is used to dynamically collect reasoning cases generated by students and integrate them into the case database after being reviewed by the instructor;

[0011] System security and privacy protection module, used to ensure data privacy and security and prevent sensitive information leakage;

[0012] The in-depth guidance module is used to guide students' reasoning process in real time through voice and dynamic clue combing functions.

[0013] Preferably, the data management module includes:

[0014] Multi-source data collector, this unit collects public data sets, historical case records and expert design cases through web crawling, API interface access and local upload. The system stores the collected data in a distributed storage system to ensure efficient data access and storage security;

[0015] The data desensitization and legality review unit uses natural language processing technology (NLP) to identify sensitive fields in the data, and removes and replaces private information through regular matching and rule engines. Legality review automatically detects data content through a preset legal and regulatory rule library;

[0016] Data labeling and classification units build criminal relationship networks based on knowledge graphs, use entity recognition technology to perform multi-dimensional classification of cases, suspects, evidence and other information, and generate indexes for rapid retrieval;

[0017] The dynamic update unit uses the scheduled task mechanism to achieve periodic updates of the data source, and combines artificial intelligence algorithms to dynamically identify data increments and update the data repository in a timely manner.

[0018] Preferably, the case generation and management module includes:

[0019] The case generation engine, based on a generative AI model, combines preset templates and data tags to generate case background, timeline, criminal behavior and other elements. After the case is generated, the system automatically verifies the logical consistency;

[0020] The case classification and retrieval unit supports classification and storage by case type, difficulty level and applicable scenario by establishing a full-text index, and provides a quick retrieval function;

[0021] Case editing and splicing tools allow instructors to modify suspect behavior, evidence chain and other fields in the case through the interface, while using modular design to splice multiple small cases into complex scenarios.

[0022] Preferably, the intelligent learning and reasoning module includes:

[0023] The AI ​​reasoning engine, based on logical reasoning algorithms, simulates the deduction logic of key nodes in the investigation process and dynamically generates prompts to guide students through the reasoning process;

[0024] Intelligent investigation assistance tools, which use image recognition, voice analysis and text mining technology to analyze case data and extract core clues, and verify whether the chain of evidence constructed by students is complete and valid;

[0025] The dynamic evaluation system records the students' reasoning path, analyzes the correctness and efficiency of reasoning based on the evaluation model, generates personalized feedback reports, and recommends supplementary learning resources.

[0026] Preferably, the user interaction module includes:

[0027] Multimodal interactive system, which supports students to interact with the system through text, voice, image, etc., and combines voice recognition and natural language processing technology to improve interaction efficiency;

[0028] Personalized learning path unit, using collaborative filtering algorithm to analyze students' historical learning behavior, dynamically adjust the difficulty of learning content and recommended cases;

[0029] Teacher-assisted management tools allow instructors to view students’ learning progress through the management interface and adjust teaching content or objectives in real time.

[0030] Preferably, the practice information collection module includes:

[0031] Case recording and generation unit, used to record the trainee's reasoning path in real time and generate case drafts;

[0032] Case submission unit, which supports students to upload new cases derived with background explanations;

[0033] The instructor review unit is used for instructors to review, optimize and score the cases submitted by students;

[0034] The case dynamic update unit is used to include the approved cases into the case database and store them by label classification;

[0035] The incentive mechanism unit encourages students to actively participate in case contributions through points, badges, and leaderboards.

[0036] Preferably, the system security and privacy protection module includes:

[0037] Privacy protection and data encryption unit, used to realize encrypted storage and transmission of data;

[0038] A hierarchical access rights management unit to limit the viewing scope of sensitive information;

[0039] Security monitoring and anomaly detection unit, used to monitor the system operation status in real time and prevent malicious operations;

[0040] The ethics and compliance management unit ensures that system functions and data usage comply with laws, regulations and ethical standards.

[0041] Preferably, the in-depth guidance module includes:

[0042] Voice interaction function, which uses natural language generation technology to provide emotional voice broadcasts and interact with students;

[0043] The clue sorting and broadcasting unit dynamically sorts out the clues obtained by students during the case reasoning process, and broadcasts the key progress and clue associations through voice;

[0044] Automatic trigger mechanism, which automatically triggers the voice guidance function when students discover new clues, unlock new stages of the case, or complete important reasoning tasks;

[0045] Process adaptability: adjust guidance content according to students’ reasoning path and learning speed to ensure guidance is targeted and consistent.

[0046] Preferably, the system supports a multi-role collaboration mode, including students, lecturers and virtual characters, to simulate the collaboration and task division between different roles in the criminal investigation process.

[0047] A method based on the above system comprises the following steps:

[0048] S1. Data processing and case generation: The data management module collects multi-source data, and after desensitization and multi-dimensional classification, transmits the data to the case generation and management module, generates dynamic cases and stores them in the case library for teaching use;

[0049] S2. Learning and reasoning interaction: Students select cases for learning through the user interaction module. The intelligent learning and reasoning module supports case analysis and logical deduction in real time. The in-depth guidance module uses voice to sort out clues and provide dynamic prompts to guide students to advance the case development.

[0050] S3. Practice and system update: The practice information collection module records the cases and reasoning results generated by the students. The instructor will include the excellent cases into the case library after review. At the same time, the system security and privacy protection module guarantees the security and legality of data circulation throughout the process, ensuring the dynamic update of teaching resources and the continuous and efficient operation of the system.

[0051] The present invention provides an AI-based intelligent crime investigation teaching system and method. It has the following beneficial effects:

[0052] 1. The present invention forms a continuous expansion mechanism for teaching resources by recording students' reasoning paths in real time, generating case drafts, and dynamically updating them to the case database after being reviewed and optimized by the lecturer through the practical information collection module. At the same time, it combines the point reward and ranking incentive functions to stimulate students' innovative thinking and enthusiasm for participation, and enhance the interactivity of the system and the practical application value of learning outcomes.

[0053] 2. The present invention uses the in-depth guidance module through voice broadcast and dynamic clue combing functions, combined with the event trigger mechanism, to provide accurate prompts and real-time support in the student's reasoning process, which not only significantly improves the logical coherence of students' analysis of complex cases, but also enhances the immersive learning experience through voice interaction, so that students can master the key nodes of case reasoning more efficiently.

[0054] 3. The present invention realizes a closed-loop process of data management, case generation, real-time reasoning and security assurance in criminal investigation teaching through multi-module collaborative work and deep technical integration, providing students with a dynamic, personalized, safe and reliable learning environment, while giving teaching resources powerful expansion capabilities and practical value, greatly improving the intelligent level of criminal investigation education. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 is a system architecture diagram of the present invention;

[0056] Figure 2 It is a schematic diagram of a data management module of the present invention;

[0057] Figure 3 A schematic diagram of a case generation and management module of the present invention;

[0058] Figure 4 It is a schematic diagram of the intelligent learning and reasoning module of the present invention;

[0059] Figure 5 is a schematic diagram of a user interaction module of the present invention;

[0060] Figure 6 It is a schematic diagram of a practical information collection module of the present invention;

[0061] Figure 7 It is a schematic diagram of the system security and privacy protection module of the present invention;

[0062] Figure 8 It is a schematic diagram of the in-depth guidance module of the present invention;

[0063] Fig. 9 It is a schematic diagram of the system operation flow of the present invention. DETAILED DESCRIPTION

[0064] The following will be combined with the drawings in the specification of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0065] Example:

[0066] Please refer to the attached Figure 1 -Attached Fig. 9 , the embodiment of the present invention provides an AI-based intelligent crime investigation teaching system, including:

[0067] The data management module collects multi-source data, performs desensitization processing, classifies and manages data, and dynamically updates data to ensure data diversity and real-time performance, providing basic support for subsequent modules. The functional implementation of this module requires multi-dimensional collaboration throughout the entire system data flow.

[0068] In this embodiment, the implementation of the multi-source data collector is as follows:

[0069] The unit builds a multi-source data collection system by calling external API interfaces, configuring web crawling rules, and supporting local data import functions. The system connects to the open data platform, receives data in JSON or CSV format, and sets crawling tasks through timer triggers. The crawled content includes case descriptions, crime trend analysis, etc. The collected data is stored in a distributed file system to ensure efficient access.

[0070] In this embodiment, the data desensitization and legality review unit is implemented as follows:

[0071] The system loads the desensitization rule library and removes or replaces sensitive information fields (such as suspect names, addresses, etc.) according to predefined rules. Text processing uses regular expressions and pre-trained NLP models to detect sensitive information in semantics and dynamically mark it. In addition, the review mechanism uses the rule engine to perform compliance checks on data content, such as checking whether it contains illegal or inappropriate content.

[0072] Sensitive information desensitization is matched by regular expressions and can be expressed as:

[0073]

[0074] Symbol Explanation

[0075] D(x): desensitized text data, x represents the original data.

[0076] x: Original data string, including case records, personal information, etc.

[0077] m i: The i-th sensitive field matched by the regular expression.

[0078] n: indicates the total number of sensitive field matches.

[0079] How to obtain:

[0080] Regular expression rules: The system presets regular rules for sensitive information.

[0081] Matching and replacement: The regular engine scans x and extracts all m that meet the rules i , and replace it with a default symbol (such as "***").

[0082] Generate desensitized data: synthesize the replaced text into D(x) and store it in the system database.

[0083] In this embodiment, the data label and classification unit are implemented as follows:

[0084] The system uses knowledge graph technology to store data in a structured manner and generate a relationship network. It extracts case elements (such as criminal methods and suspect identities) through entity recognition algorithms, classifies them into layers according to dimensions such as case complexity and legal provisions, and generates a searchable tag index.

[0085] In this embodiment, the dynamic update unit is implemented as follows:

[0086] The system uses an event trigger mechanism to periodically call the crawler to obtain the latest data. The new data is compared with the existing database content, and the incremental part is updated to ensure the real-time nature of the case library.

[0087] The case generation and management module is responsible for dynamically generating and efficiently managing cases. Its implementation technologies include generative AI models, classification retrieval algorithms, and visual editing tools.

[0088] In this embodiment, the implementation of the case generation engine is:

[0089] The system automatically generates case background, criminal methods, time clues, etc. by loading pre-trained generative artificial intelligence models and customized templates. The generation process combines the data labels of the aforementioned knowledge graph, automatically screens relevant elements, and applies rules to constrain the logical consistency of the model-generated content.

[0090] In this embodiment, the case classification and retrieval unit is implemented as follows:

[0091] The system uses full-text search technology to build an index and classifies and stores cases by field identifiers (such as case difficulty and applicable disciplines). The search function uses an inverted index structure to achieve fast keyword matching and provide efficient query results.

[0092] In this embodiment, the case editing and splicing tool is implemented as follows:

[0093] Provides a graphical interface that supports instructors to adjust case content through drag-and-drop operations. The system has a built-in modular design that allows users to splice multiple small cases into complex scenarios according to timelines or logical relationships to generate a multi-stage case.

[0094] The intelligent learning and reasoning module supports students' learning and reasoning practice through reasoning engines and dynamic feedback.

[0095] In this embodiment, the implementation of the AI ​​reasoning engine is:

[0096] This unit uses logistic regression and Bayesian network models to simulate the logic path of criminal investigation. After the inference engine loads the case data, it generates key nodes such as suspects and evidence chains, and automates the inference process through weight allocation and dynamic adjustment;

[0097] The calculation formula is:

[0098]

[0099] Symbol explanation:

[0100] P(H|E): The posterior probability of hypothesis H (e.g., the suspect is guilty) given the known evidence E.

[0101] P(H): The prior probability of hypothesis H, derived from historical data.

[0102] P(E|H): The conditional probability of evidence E appearing when hypothesis H is true, which is derived from case data statistics.

[0103] P(E): The marginal probability of evidence E occurring, obtained by weighted summation of the conditional probabilities of all hypotheses.

[0104] How to obtain:

[0105] Prior probability calculation: Initialize the probability distribution of P(H) based on the statistical results of historical cases.

[0106] Conditional probability training: Train the Bayesian network structure through case data to obtain the conditional probability table of P(E|H).

[0107] Real-time reasoning: After the student selects or adds evidence E, the reasoning engine calculates P(H|E) in real time and provides probability-based reasoning results.

[0108] In this embodiment, the intelligent investigation auxiliary tool is implemented as follows:

[0109] The system loads deep learning models (such as YOLO) to analyze behavioral patterns in surveillance videos, extracts text data with OCR technology, and verifies the integrity of the chain of evidence. The system uses API calls to the audio processing module to analyze phone records and automatically generate semantic summaries.

[0110] In this embodiment, the implementation of the dynamic evaluation system is as follows:

[0111] The system records the trainees’ reasoning paths, analyzes the accuracy of decisions and the efficiency of reasoning, and generates rule-based feedback reports. At the same time, it recommends knowledge modules related to weak links to help trainees optimize their reasoning logic.

[0112] The user interaction module supports the interaction between students and the system through multimodal interaction.

[0113] In this embodiment, the multimodal interaction system is implemented as follows:

[0114] The unit uses speech recognition, text analysis and image processing technologies to achieve multiple input methods. The system calls the STT (speech to text) model to process the trainee's voice commands, and also supports uploading case scene pictures through images to generate relevant clues.

[0115] In this embodiment, the personalized learning path unit is implemented as follows:

[0116] The system uses collaborative filtering algorithms to dynamically adjust learning content based on students’ historical data. The algorithm considers two dimensions, completion and error rate, in case recommendations to customize a learning path of appropriate difficulty for students.

[0117] In this embodiment, the implementation of the teacher-assisted management tool is as follows:

[0118] The management interface displays students’ real-time learning progress, including the number of completed cases and the quality of feedback. Instructors can adjust learning goals or allocate additional resources based on the suggestions generated by the system.

[0119] The in-depth guidance module supports students through voice prompts and clue combing functions.

[0120] In this embodiment, the voice interaction function is implemented as follows:

[0121] The speech generation module uses TTS (text-to-speech) technology to convert prompts into speech broadcasts in real time. The system loads multiple emotion expression models, and the generated speech is synchronized with the trainee's reasoning process.

[0122] In this embodiment, the clue combing and broadcasting unit is implemented as follows:

[0123] The system combines knowledge graphs to logically sort out the clues that students have obtained. Key nodes are dynamically extracted through semantic generation models, and voice prompts are generated in real time.

[0124] The knowledge graph sorts out the logical relationship between clues and uses the adjacency matrix to represent the graph structure:

[0125]

[0126] Symbol explanation:

[0127] A[i][j]: element in the adjacency matrix, representing node v i and v j Whether there is a relationship.

[0128] v i , v j : A node in the knowledge graph, representing a case element (such as a suspect, a time point, or evidence).

[0129] 1: Indicates that there is a direct association between nodes.

[0130] 0: Indicates that there is no direct association between nodes.

[0131] How to obtain:

[0132] Node generation: Extract graph nodes based on case information, including suspects, evidence, case background, etc.

[0133] Edge relationship definition: Generate edge relationships based on the logical connection between evidence (such as "evidence points to the suspect") and construct the adjacency matrix A[i][j].

[0134] Logical path calculation: Use graph traversal algorithms (such as DFS or BFS) to find the associated paths between clues and generate dynamic prompts.

[0135] In this embodiment, the automatic trigger mechanism is implemented as follows:

[0136] The system uses event detection algorithms to automatically activate the guidance function when students unlock new stages or complete important tasks, prompting the next steps.

[0137] The practice information collection module dynamically records the students' reasoning process and supports case updates.

[0138] In this embodiment, the case recording and generating unit is implemented as follows:

[0139] The system generates operation logs by monitoring the students' reasoning process, extracts key reasoning nodes and evidence chains, and automatically generates case drafts.

[0140] In this embodiment, the implementation of the case submission unit is:

[0141] Students can upload custom cases through the visual interface, attach background explanations and logical analysis. The system automatically verifies the cases to detect whether there are logical conflicts.

[0142] In this embodiment, the lecturer review unit is implemented as follows:

[0143] The instructor interface displays case content, scoring, and optimization suggestion fields, and approved cases are directly pushed to the case library.

[0144] In this embodiment, the case dynamic update unit is implemented as follows:

[0145] The system manages new cases through labeling, updates the approved cases to the case library according to classification dimensions, and synchronizes them to the search engine.

[0146] The system security and privacy protection module runs through the entire teaching system to ensure data security, access compliance, and reliability of interaction between modules. The design of this module integrates encryption technology, access control, and real-time monitoring mechanisms to effectively ensure the storage and transmission security of sensitive data.

[0147] This module mainly provides comprehensive security protection for the operation of the system through four functional units: data encryption technology, hierarchical access rights management, security monitoring, and ethical compliance review. During the interaction between modules, the module monitors the data flow in real time and intercepts and processes abnormal behaviors. Through multi-level security strategies, this module achieves comprehensive protection of data privacy, system operation, and user behavior.

[0148] In this embodiment, the implementation of the privacy protection and data encryption unit

[0149] High-strength encryption is used during data storage and transmission. The system uses symmetric encryption algorithms (such as AES-256) to encrypt sensitive fields in the database to ensure that data cannot be directly read when stored.

[0150] During data transmission, an encrypted communication channel is established through TLS (Transport Layer Security Protocol) to ensure that data exchange between modules is protected from eavesdropping and tampering.

[0151] The system's data encryption operation is combined with a key management mechanism and uses distributed key storage to prevent the risk of single-point key leakage.

[0152] In this embodiment, the hierarchical access rights management unit is implemented as follows:

[0153] The system assigns different access rights based on user roles (such as students, instructors, and administrators) to ensure that sensitive information can only be viewed and operated by authorized personnel.

[0154] The permission management strategy is based on the RBAC (role-based access control) model. The user's access rights are dynamically bound to roles and can be adjusted according to needs.

[0155] The permission allocation expression of the RBAC (role-based access control) model in the system is as follows:

[0156]

[0157] Symbol explanation:

[0158] P(u,r,o,a): indicates the permission of user u to perform operation a on object o. The output is 1 for permission and 0 for rejection.

[0159] u: User, refers to a student, instructor or administrator.

[0160] r: Role, bound to a user, such as a student role, instructor role, or administrator role.

[0161] o: Object, resources or data in the system, such as case libraries, log records, learning reports, etc.

[0162] a: Operation, the operation the user wants to perform, such as "view", "modify" or "delete".

[0163] UR: A set of user-role mapping relationships, defining which users belong to which roles. (u,r)∈UR means that user u is assigned role r.

[0164] PA: permission allocation set, which defines the role's operation permissions on objects. (r,o,a)∈PA means that role r has the permission to operate a on object o.

[0165] User role assignment: Initialize the UR set through user registration information and role configuration strategy.

[0166] Permission rule configuration: The administrator predefines the PA set based on system design requirements, and the specific permissions are stored in the permission management database.

[0167] Dynamic permission verification: When a user requests access to a resource, the system determines whether the permission is met based on P(u, r, o, a).

[0168] The system supports fine-grained permission configuration. For example, lecturers can view students' learning data but cannot access the core case database, while administrators have management permissions for the global configuration of the system.

[0169] In this embodiment, the security monitoring and anomaly detection unit is implemented as follows:

[0170] The system deploys a real-time security monitoring module to monitor the system operation status, including user behavior logs, data access records and inter-module communication status.

[0171] Anomaly detection uses machine learning models to analyze user behavior patterns and identify abnormal operations, such as unauthorized access attempts, large data downloads, or illegal modification requests;

[0172] Abnormal behavior is detected using user behavior feature vectors, and the behavior deviation is calculated using the Euclidean distance formula:

[0173]

[0174] Symbol explanation:

[0175] d(u i ,u m ): User u i Current behavior and normal behavior model u m The larger the value, the higher the probability of abnormality.

[0176] x ij :User u i The value of the j-th behavior feature.

[0177] x mj : The mean of the normal behavior model on the jth behavior feature.

[0178] n: The total dimension of behavioral characteristics, such as access frequency, request size, etc.

[0179] How to obtain:

[0180] Feature extraction: parse user behavior logs (such as request records, access time, etc.) to extract behavior features x ij .

[0181] Normal model generation: Calculate the mean x of each feature through historical data mj and standard deviation to establish a normal behavior model.

[0182] Deviation calculation: Real-time calculation of u i The Euclidean distance d(u i ,u m )>T(preset threshold T), an abnormal alarm is triggered.

[0183] The system will trigger an alarm mechanism for detected abnormal behavior, notify the administrator in real time, and perform blocking or isolation operations according to the policy.

[0184] In this embodiment, the implementation of the ethics and compliance management unit is:

[0185] The system loads the ethical rule base and legal and regulatory rule set, and uses the rule engine to conduct compliance reviews on new cases, content uploaded by students, and all data processed by the system.

[0186] The system automatically analyzes the case content to identify whether it contains sensitive information or plots that violate ethical standards, such as overly detailed descriptions of violence or data that violates privacy.

[0187] The compliance review module supports dynamic updating of the rule base to adapt to new laws, regulations and ethical requirements.

[0188] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. The AI-based intelligent crime investigation teaching system is characterized by: include: Data management module, used to collect, process and manage crime investigation related data to ensure data diversity and quality; Case generation and management module, which is used to dynamically generate and manage teaching cases based on artificial intelligence, including case background, clue relationships and reasoning paths; Intelligent learning and reasoning module, used to support students in learning criminal investigation knowledge, carrying out case reasoning and practical operations; User interaction module, which is used to provide multimodal interaction and support personalized learning paths and dynamic feedback; The practical information collection module is used to dynamically collect reasoning cases generated by students and integrate them into the case database after being reviewed by the instructor; System security and privacy protection module, used to ensure data privacy and security and prevent sensitive information leakage; The in-depth guidance module is used to guide students' reasoning process in real time through voice and dynamic clue combing functions.

2. The AI-based intelligent crime investigation teaching system according to claim 1 is characterized in that: The data management module comprises: Multi-source data collector for acquiring data from public datasets, historical case records, expert design cases, and web crawling; Data desensitization and legality review unit, used to remove sensitive information and ensure that data complies with legal and ethical requirements; Data labels and classification units, based on knowledge graph technology to manage data in multiple dimensions; Dynamic update unit to regularly introduce new data reflecting the latest crime trends.

3. The AI-based intelligent crime investigation teaching system according to claim 1 is characterized in that: The case generation and management module includes: A case generation engine that dynamically generates case background, suspect behavior, time clues, and reasoning paths based on generative AI; Case classification and retrieval unit, which supports classification and quick retrieval of cases by difficulty, type and applicable scenarios; Case editing and splicing tools are used to combine multiple small cases into complex scenarios and allow instructors to adjust the cases.

4. The AI-based intelligent crime investigation teaching system according to claim 1 is characterized in that: The intelligent learning and reasoning module includes: AI reasoning engine, used to simulate the criminal investigation reasoning process and support dynamic case deduction; Intelligent investigation assistance tools to analyze phone records, surveillance videos and transaction data in cases and verify students’ chain of evidence; A dynamic assessment system that records students' reasoning paths, generates personalized feedback reports and recommends relevant knowledge.

5. The AI-based intelligent crime investigation teaching system according to claim 1 is characterized in that: The user interaction module comprises: Multimodal interaction system, supporting multiple interaction modes such as text, voice, image and video; Personalized learning path unit, used to dynamically adjust learning content based on students’ learning data; Teacher-assisted management tool for instructors to monitor students' progress in real time and provide teaching suggestions.

6. The AI-based intelligent crime investigation teaching system according to claim 1 is characterized in that: The practice information collection module includes: Case recording and generation unit, used to record the trainee's reasoning path in real time and generate case drafts; Case submission unit, which supports students to upload new cases derived with background explanations; The instructor review unit is used for instructors to review, optimize and score the cases submitted by students; The case dynamic update unit is used to include the approved cases into the case database and store them by label classification; The incentive mechanism unit encourages students to actively participate in case contributions through points, badges, and leaderboards.

7. The AI-based intelligent crime investigation teaching system according to claim 1 is characterized in that: The system security and privacy protection module includes: Privacy protection and data encryption unit, used to realize encrypted storage and transmission of data; A hierarchical access rights management unit to limit the viewing scope of sensitive information; Security monitoring and anomaly detection unit, used to monitor the system operation status in real time and prevent malicious operations; The ethics and compliance management unit ensures that system functions and data usage comply with laws, regulations and ethical standards.

8. The AI-based intelligent crime investigation teaching system according to claim 1 is characterized in that: The in-depth guidance module includes: Voice interaction function, which uses natural language generation technology to provide emotional voice broadcasts and interact with students; The clue sorting and broadcasting unit dynamically sorts out the clues obtained by students during the case reasoning process, and broadcasts the key progress and clue associations through voice; Automatic trigger mechanism, which automatically triggers the voice guidance function when students discover new clues, unlock new stages of the case, or complete important reasoning tasks; Process adaptability: adjust guidance content according to students’ reasoning path and learning speed to ensure guidance is targeted and consistent.

9. The AI-based intelligent crime investigation teaching system according to claim 1 is characterized in that: The system supports a multi-role collaboration mode, including students, instructors and virtual characters, simulating the collaboration and task division between different roles in the criminal investigation process.

10. A method based on the system of claim 1, characterized in that: The following steps are involved: S1. Data processing and case generation: The data management module collects multi-source data, and after desensitization and multi-dimensional classification, transmits the data to the case generation and management module, generates dynamic cases and stores them in the case library for teaching use; S2. Learning and reasoning interaction: Students select cases for learning through the user interaction module. The intelligent learning and reasoning module supports case analysis and logical deduction in real time. The in-depth guidance module uses voice to sort out clues and provide dynamic prompts to guide students to advance the case development. S3. Practice and system update: The practice information collection module records the cases and reasoning results generated by the students. The instructor will include the excellent cases into the case library after review. At the same time, the system security and privacy protection module guarantees the security and legality of data circulation throughout the process, ensuring the dynamic update of teaching resources and the continuous and efficient operation of the system.

Images