Educational Assessment Method and Device Based on AIGC Technology Interactive Video

Abstract

This application relates to the field of video production technology, and provides an educational assessment method and apparatus for interactive videos based on AIGC technology. The method, based on the current user's information and the type of cybersecurity to be assessed, searches a stored set of cybersecurity incident cases for current cybersecurity incident cases and corresponding learning objectives, and uses AIGC technology to generate a narrative video; it obtains the target narrative selection control triggered by the current user and the decision score for triggering the target narrative selection control; it then finds target cybersecurity incident cases that satisfy the narrative information corresponding to the narrative video; it uses the target cybersecurity incident cases as new current cybersecurity incident cases to generate narrative videos, until the current user's viewing time reaches a preset duration; and it determines the current user's security awareness level based on the decision scores corresponding to the narrative videos watched. This method improves user learning effectiveness and participation through interactive experience.

Description

Technical Field

[0001] This application relates to the field of video production technology, and more specifically, to an educational assessment method and apparatus for interactive videos based on AIGC technology. Background Technology

[0002] Currently, cybersecurity education and training generally employ traditional teaching methods, such as lectures, training courses, and manual reading, as well as some digital tools, such as online tutorials, mock tests, and animated demonstrations. While these methods have improved users' cybersecurity awareness to some extent, they often suffer from insufficient interactivity, limited experiential learning, and difficulty in sustaining learning outcomes. Summary of the Invention

[0003] The purpose of this application is to provide an educational assessment method and apparatus based on interactive video using AIGC technology, which solves the above-mentioned problems existing in the prior art and improves the user's learning effect and participation through interactive experience.

[0004] Firstly, an educational assessment method based on interactive video using AIGC technology is provided, applicable to cybersecurity education systems. This method may include:

[0005] Based on the current user's information and the network security type to be evaluated, the system searches the stored network security incident case set for the current network security incident case and the corresponding teaching objective; the network security incident case set contains network security incident cases of different network security types.

[0006] Based on the current cybersecurity incident case and the corresponding teaching objectives, AIGC technology is used to generate a narrative video; the video frame at the end of the narrative video includes a narrative selection control that can trigger different storylines;

[0007] Obtain the target story selection control triggered by the current user and the decision score that triggered the target story selection control; the decision score is a score used to characterize the level of safety awareness, determined based on the teaching objectives corresponding to the story video and the storyline of the target story selection control.

[0008] Based on the plot direction corresponding to the target plot selection control, find the target cybersecurity incident cases in the cybersecurity incident case set that satisfy the plot information corresponding to the plot video;

[0009] The target cybersecurity incident case is used as a new current cybersecurity incident case, and the execution steps are returned: Based on the current cybersecurity incident case and the corresponding teaching objectives, a narrative video is generated until the current user's viewing time reaches the preset duration;

[0010] Based on the correspondence between different configuration decision scores and different security awareness levels, and the decision scores corresponding to the plot videos watched by the current user, the security awareness level of the current user is determined.

[0011] In an optional implementation, the user information includes a user identifier and user attribute information;

[0012] Find any network security incident case and corresponding teaching objective in the stored set of network security incident cases that meet the network security type to be evaluated and the user information, including:

[0013] Find network security incident cases in the stored network security incident case collection that meet the network security type to be evaluated;

[0014] If the stored mapping table of user identifiers and security awareness level values ​​contains the user identifier and the corresponding security awareness level value, then select the current network security incident case and the corresponding teaching objective that meet the security awareness level value from the found network security incident cases.

[0015] If the user identifier is not included in the mapping table between the stored user identifier and the security awareness level value, the initial security awareness level value of the current user is determined according to the user attribute information, and any network security event case that meets the initial security awareness level value and the corresponding teaching objective are selected from the network security event cases found.

[0016] In an optional implementation, the plot information includes a plot summary and character information;

[0017] Based on the aforementioned current cybersecurity incident cases and corresponding teaching objectives, AIGC technology is used to generate narrative videos, including:

[0018] The trained sequence generation model is used to process the network security incident cases and corresponding teaching objectives to obtain plot outlines and character information that meet the teaching objectives.

[0019] Using a pre-defined AIGC technology, the plot synopsis and character information are combined to generate script text for the cybersecurity incident type;

[0020] A pre-defined conditional generative adversarial network is used to process the script text to obtain the corresponding plot video.

[0021] In an optional implementation, a pre-defined conditional generative adversarial network is used to process the script text to obtain a corresponding plot video, including:

[0022] The generator network in the conditional generative adversarial network generates a sequence of video frames based on the input script text; the video information in each video frame includes scene, character, and action.

[0023] The discriminator network in the conditional generative adversarial network distinguishes the scenes, characters, and actions in each video frame from the scenes, characters, and actions in the real world.

[0024] If the discrimination result meets the preset error condition, then the video frame sequence is determined as the plot video corresponding to the script text.

[0025] In an optional implementation, the security awareness level of the current user is determined based on the correspondence between different configuration decision scores and different security awareness levels, and the decision scores corresponding to the drama videos watched by the current user, including:

[0026] If the current user's continuous viewing time reaches a preset viewing time threshold, then obtain the decision score corresponding to the plot video watched by the current user;

[0027] The current user's decision score is determined by summing the decision scores of the viewed story videos;

[0028] Based on the correspondence between different configured decision scores and different security awareness levels, the security awareness level corresponding to the current user's decision score is determined.

[0029] Store the correspondence between the current user's user ID and the corresponding security awareness level value.

[0030] In an optional implementation, based on the plot development corresponding to the target plot selection control, different target cybersecurity incident cases in the cybersecurity incident case set that satisfy the plot information corresponding to the plot video are found, including:

[0031] Using the established deep learning natural language processing model, the plot direction corresponding to the target plot selection control, the plot information of the previous plot video of the current user, and the character information are analyzed to obtain plot analysis results;

[0032] Based on the plot analysis results, different target cybersecurity incident cases that satisfy the plot information are found in the set of cybersecurity incident cases.

[0033] In an optional implementation, the method further includes:

[0034] Based on the user's action of selecting the target storyline control, obtain the user's behavior data;

[0035] By performing in-depth analysis on the behavioral data, the learning characteristics of the current user's viewing of the plot video are obtained.

[0036] Based on the learning characteristics, learning suggestions are provided to the current user.

[0037] Secondly, an interactive video-based educational assessment device based on AIGC technology is provided for application in a cybersecurity education system. This device may include:

[0038] The search unit is used to search for current network security incident cases and corresponding teaching objectives in the stored network security incident case set based on the current user's user information and the network security type to be evaluated; the network security incident case set contains network security incident cases of different network security types;

[0039] The generation unit is used to generate a narrative video based on the current cybersecurity incident case and the corresponding teaching objectives, using AIGC technology; the video frame at the end of the narrative video includes a narrative selection control that can trigger different narratives;

[0040] The acquisition unit is used to acquire the target plot selection control triggered by the current user and the decision score that triggered the target plot selection control; the decision score is a score used to characterize the level of safety awareness, which is determined based on the teaching objectives corresponding to the plot video and the plot direction of the target plot selection control.

[0041] The search unit is also used to find target cybersecurity incident cases in the cybersecurity incident case set that satisfy the plot information corresponding to the plot video, based on the plot direction corresponding to the target plot selection control;

[0042] The triggering unit is used to take the target network security incident case as a new current network security incident case and trigger the generation unit to execute: based on the current network security incident case and the corresponding teaching objectives, generate a story video until the current user's viewing time reaches the preset time.

[0043] The determining unit is used to determine the current user's security awareness level based on the correspondence between different configured decision scores and different security awareness levels, and the decision scores corresponding to the plot videos watched by the current user.

[0044] Thirdly, an electronic device is provided, which includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus;

[0045] Memory, used to store computer programs;

[0046] When a processor executes a program stored in memory, it implements any of the steps described in the first aspect above.

[0047] Fourthly, a computer-readable storage medium is provided, wherein a computer program is stored therein, and when executed by a processor, the computer program implements the steps of any of the methods described in the first aspect above.

[0048] The interactive video-based educational assessment method provided in this application, based on AIGC technology, can search for current cybersecurity incident cases and corresponding teaching objectives in a stored set of cybersecurity incident cases, based on the current user's user information and the cybersecurity type to be assessed. The set of cybersecurity incident cases includes cybersecurity incident cases of different cybersecurity types. Based on the current cybersecurity incident cases and corresponding teaching objectives, AIGC technology is used to generate a narrative video. The video frame at the end of the narrative video includes different triggerable narrative selection controls. The method obtains the target narrative selection control triggered by the current user and the decision score for triggering the target narrative selection control. The decision score is a score used to characterize the level of security awareness, determined based on the teaching objectives corresponding to the narrative video and the narrative direction of the target narrative selection control. Based on the narrative direction corresponding to the target narrative selection control, the method finds target cybersecurity incident cases in the set of cybersecurity incident cases that satisfy the narrative information corresponding to the narrative video. The target cybersecurity incident case is used as the new current cybersecurity incident case, and the execution steps are returned: generating a narrative video based on the current cybersecurity incident case and corresponding teaching objectives until the current user's viewing time reaches a preset duration; determining the current user's security awareness level based on the configured correspondence between different decision scores and different security awareness levels, and the decision scores corresponding to the narrative videos currently viewed by the current user. This method improves user learning outcomes and engagement through interactive experiences. Attached Figure Description

[0049] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0050] Figure 1 A flowchart illustrating an educational assessment method based on interactive video using AIGC technology, provided as an embodiment of this application;

[0051] Figure 2 A schematic diagram of the structure of an interactive video-based educational assessment device based on AIGC technology provided in this application embodiment;

[0052] Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0053] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0054] The interactive video-based educational assessment method based on AIGC technology provided in this application can be applied to a terminal or to a communication system that connects a server and a terminal.

[0055] If applied to a terminal, the user interacts through the terminal's interface, and the processor within the terminal executes the method steps of the interactive video educational assessment method based on AIGC technology. If applied to a communication system where a server and a terminal communicate, the user interacts through the terminal's interface, and the processor within the server executes the method steps of the interactive video educational assessment method based on AIGC technology. The specific division of labor for the work steps can be determined according to actual needs, and this application does not impose any restrictions on it.

[0056] The server can be a physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms. The terminal can be a user equipment (UE) such as a mobile phone, smartphone, laptop, digital broadcast receiver, personal digital assistant (PDA), tablet computer (PAD), handheld device, in-vehicle device, wearable device, computing device or other processing device connected to a wireless modem, mobile station (MS), mobile terminal, etc.

[0057] The implementation process of the interactive video-based educational assessment method based on AIGC technology provided in this application covers several key steps and technology integrations to ensure that the system can effectively generate and deliver high-quality, personalized, and up-to-date cybersecurity awareness education content. Before implementing this application, a basic educational framework and case materials need to be established, including:

[0058] Regarding the establishment of a basic education framework:

[0059] When designing a cybersecurity awareness education system, the primary task is to create a comprehensive foundational education framework. This framework is based on in-depth educational theoretical research and practical application, ensuring the scientific rigor and practicality of its content. Specific steps include:

[0060] 1) Knowledge domain division: The content of cybersecurity awareness education is divided into different knowledge modules, including password security, prevention of cyber fraud, data privacy protection, protection against malware, cyber law and ethics, etc.

[0061] 2) Hierarchical design: The knowledge points are arranged in a progressive manner from beginner to advanced levels to ensure that learners of different levels can find a suitable starting point. For example, the basic knowledge section introduces the basic security concepts of the network and risk identification methods; the intermediate stage explores specific security operations and strategy formulation; and the advanced stage involves enterprise-level security management and emergency response strategies.

[0062] 3) Establishment of standards and specifications: Based on the latest domestic and international cybersecurity standards, regulations, and best practices, determine the standard teaching content and learning objectives for each knowledge point.

[0063] 4) Interactivity and Practicality: Design interactive learning activities and simulation exercises to enable trainees to consolidate their knowledge and improve their practical skills through practice.

[0064] Regarding obtaining case materials:

[0065] To enhance teaching effectiveness, it is necessary to collect a large number of real-world cybersecurity incidents as a case study database, i.e., to construct a collection of cybersecurity incident cases. These cases can be used for:

[0066] 1) Scenario Reenactment: By recreating case scenarios, trainees can understand the true nature of cyber threats and improve their awareness and sensitivity to potential risks.

[0067] 2) Problem Analysis: Analyze the causes, evolution, and consequences of the problems in the case, guide trainees to think about how to prevent similar situations from happening, and learn how to deal with them.

[0068] 3) AI Algorithm Training and Content Generation: Using these case data to train AI models, AI can intelligently generate targeted teaching content, test questions, or simulation exercises, providing personalized services to each user and further enhancing the educational experience and effectiveness.

[0069] The preferred embodiments of this application are described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit this application. Furthermore, the embodiments and features in the embodiments of this application can be combined with each other without conflict.

[0070] Figure 1 This is a flowchart illustrating an educational assessment method based on interactive video using AIGC technology, provided as an embodiment of this application. Figure 1 As shown, the method may include:

[0071] Step S110: Based on the current user's user information and the network security type to be evaluated, find the current network security event case and the corresponding teaching objective in the stored network security event case set.

[0072] The cybersecurity incident case collection includes cybersecurity incidents of various types. The attribute information for each case can include the attack type, attack method, attacker, victim, scope of impact, countermeasures, and key points for security education. User information can include user identification and attributes such as age and occupation. The collection covers core knowledge points including various threat types (such as phishing, malware, social engineering, and data breaches), protective measures, regulations and standards, and best practices.

[0073] In one example, a set of cybersecurity incident cases can be stored in the form of a knowledge graph, where nodes in the knowledge graph represent cybersecurity incident cases and edges represent the relationships between two nodes (i.e., cybersecurity incident cases). It can be understood that the length of the edge can represent the degree of relationship between two connected nodes. For example, the longer the edge, the greater the degree of relationship between the two nodes, and vice versa.

[0074] Specifically, the knowledge graph involved in this application is briefly introduced below:

[0075] 1. Entities (Nodes)

[0076] a. Case Studies: Each specific cybersecurity incident is treated as a separate node, including basic information such as its name, number, time of occurrence, and the industry or company involved.

[0077] b. Attack types: such as DDoS attacks, APT attacks, ransomware attacks, data breaches, phishing attacks, supply chain attacks, etc.

[0078] c. Attack methods: such as vulnerability exploitation, social engineering, malware, man-in-the-middle attacks, denial-of-service attacks, and other specific implementation methods.

[0079] d. Victims: This includes affected businesses, organizations, individuals, or systems, as well as their industry attributes, geographical location, etc.

[0080] e. Attackers: such as known hacker groups, state-sponsored attackers, independent criminals, etc., and their motives and background information.

[0081] f. Scope of impact: This includes quantitative indicators such as the number of affected systems, the amount of data affected, economic losses, and business interruption time.

[0082] g. Response measures: such as emergency response procedures, patch application, system hardening, legal accountability, and other specific response actions and results.

[0083] 2. Edges

[0084] a. Event-Attack Type: Indicates the type of attack a given event belongs to.

[0085] b. Event-Attack Methods: Indicates the specific attack methods used in the event.

[0086] c. Event-Victim: This refers to the specific entity that was attacked in the event.

[0087] d. Event-Attacker: Indicates the person or entity behind the event.

[0088] e. Event - Scope of Impact: Describe the specific losses and impacts caused by the event.

[0089] f. Incident - Response: Record the response strategies and results taken in response to the incident.

[0090] g. Attack Type - Related Events: Links multiple events that occur under the same type.

[0091] h. Attack Methods - Related Events: Show other events that use the same methods.

[0092] i. Victims-Related Incidents: List all cybersecurity incidents that have occurred to this entity.

[0093] j. Attacker-related events: A summary of all attacks initiated by the same attacker.

[0094] 3. Structure

[0095] a. Center-radial structure: Taking the event case as the central node, it radiates outwards to relate to the attack type, attack method, victim, attacker, scope of impact, and countermeasures.

[0096] b. Hierarchical structure: Organize in layers according to event level (e.g., national level, industry level, enterprise level), attack severity (e.g., high, medium, low), or timeline (e.g., annual, quarterly).

[0097] c. Topic Clustering: Related incident cases are clustered based on factors such as attack methods, industry sectors, and geographical regions to form topic clusters, which facilitates the rapid identification of cases of specific types or situations.

[0098] Returning to step S110, in specific implementation, at least one network security event case that meets the network security type to be evaluated is found in the stored set of network security event cases;

[0099] If the stored mapping table of user identifiers and security awareness level values ​​contains user identifiers and corresponding security awareness level values, then from at least one network security incident case found, select the current network security incident case that meets the security awareness level value and the corresponding teaching objective.

[0100] If the stored mapping table of user identifiers and security awareness levels does not contain a user identifier (i.e., the current user is an initial user), then the initial security awareness level value for the current user is determined based on the user attribute information. For example, the initial security awareness level value for a 6-year-old student is 1, and the initial security awareness level value for a 30-year-old student is 30. A higher security awareness level value indicates a stronger level of security awareness. Then, from the retrieved cybersecurity incident cases, any cybersecurity incident case that meets this initial security awareness level value and its corresponding teaching objective are selected.

[0101] Step S120: Based on current cybersecurity incident cases and corresponding teaching objectives, generate narrative videos using AIGC technology.

[0102] The video frame at the end of the story video can include a story selection control that can trigger different storylines.

[0103] In practice, a pre-trained sequence generation model (such as Transformer) is used to process cybersecurity incident cases and corresponding teaching objectives to obtain plot information that meets the teaching objectives. This plot information can include a plot summary and character information; character information can include the number of characters, their personality traits, their appearance, and / or their profession. Specifically, the encoder of the pre-trained sequence generation model is responsible for understanding the content of the cybersecurity incident cases, while the decoder generates the plot line, i.e., the plot information, based on the teaching objectives and the content understood by the encoder.

[0104] Specifically, the Transformer model with self-attention mechanism is trained by maximizing likelihood estimation (MLE) for a given set of cybersecurity incident cases and the corresponding teaching target vectors for those cases.

[0105]

[0106] in, This means we want to find the maximum value of the model parameter θ such that the following expression is maximized. (c i ,t i ) ~D represents the mean (expected value) of all context-target pairs sampled according to the data distribution D. logP θ (t i |c i G) is the model's performance on target t given parameters θ. i In context c i The logarithm of the conditional probability given the global variable G.

[0107] The loss function is designed to minimize the difference between the teaching content and teaching objectives reflected in the generated storyline, while ensuring the rationality and innovation of the storyline.

[0108] Then, using pre-defined AIGC technology, the plot synopsis and character information are combined to generate script text for cybersecurity incidents. Specifically, based on a set of cybersecurity incident cases, an AI algorithm model trained using AIGC technology can be used to process the plot synopsis and character information, generating script text that includes dialogue, action descriptions, emotional expressions, and plot choice points (i.e., plot turning points) that can trigger different plots. The AI ​​algorithm model can learn from a large number of video script samples, understand elements such as script structure, character dialogue, plot turning points, and thematic expression, and can generate coherent and creative script text based on given inputs such as the theme, plot setting, and character settings.

[0109] In one example, plot generation algorithms, based on models such as reinforcement learning (RL) or variational autoencoders (VAE), automate the planning of plot development. These algorithms can explore possible storylines and generate reasonable plot twists and endings based on set initial conditions and target states. Character dialogue generation utilizes pre-trained dialogue generation models, such as BERT and DialoGPT, to simulate natural dialogue between characters, ensuring that the dialogue content is both consistent with the characters' personalities and advances the plot.

[0110] Next, a pre-defined conditional generative adversarial network (GAN) is used to process the script text, resulting in a corresponding video clip. Specifically, the generator network in the GAN generates a sequence of video frames (i.e., videos) based on the input script text. Each video frame contains information such as scene, characters, and actions. The discriminator network in the GAN compares the scene, characters, and actions in each video frame with real-world scenes, characters, and actions. If the comparison result meets a pre-defined error condition, the video frame sequence is identified as the corresponding video clip. If the comparison result does not meet the pre-defined error condition, the parameters of the generator network are adjusted to obtain a comparison result that satisfies the pre-defined error condition.

[0111] In Conditional Generative Adversarial Networks (cGANs), the generator `G(z,c)` receives a random noise vector `z` and conditional information (such as text description) `c` to generate visual content, while the discriminator `D(x,c)` judges whether the input visual content `x` is real. The loss function consists of two parts:

[0112]

[0113] In the formula, the value function V(D,G) is expressed as:

[0114]

[0115] In this expression: This represents the expected log probability that the discriminator correctly identifies the real sample and gives a high score (probability close to 1) when x is sampled from the real data distribution pdata(x). This indicates that when z is sampled from the prior distribution p z When the generator G maps the sample G(z) to the latent space and generates samples G(z,c), the discriminator correctly judges these generated samples as non-real samples and gives them a low score (probability close to 0) expected log probability.

[0116] The process of automatically generating video scripts, visuals, and sound elements using AIGC technology involves a complex stack of technologies and key algorithm models. The following is a detailed description of the generation process for these elements:

[0117] I. Automatic Generation of Video Scripts

[0118] - Script outline generation: Input video type, theme, basic plot framework and other information, and the AIGC system uses plot generation algorithm to output script outline, including key nodes such as main events, character relationships, and climax of conflict.

[0119] - Detailed script writing: Based on the outline, the NLP model fills in the details according to the standard format of video scripts (such as scene descriptions, character actions, dialogues, etc.) to generate a complete script text.

[0120] - Dialogue optimization and review: The generated character dialogues may need to be further optimized through means such as sentiment analysis and logical consistency checks to ensure that the dialogues are consistent with both the character settings and the overall logic of the script.

[0121] II. Screen Generation:

[0122] -Scene Design: Based on the script description, use CV models to automatically generate or select suitable background materials to construct scene layouts, including indoor, outdoor, urban, and natural landscapes.

[0123] -Character Modeling and Animation: Based on the character design, generate or select the corresponding 3D model, and use motion generation model to give the character actions and expressions that conform to the plot.

[0124] - Lighting and Special Effects: Utilize AI technology to simulate real-world lighting effects and add necessary visual effects (such as smoke, fire, magic, etc.) to enhance the visual appeal.

[0125] - Rendering output: Integrates elements such as scenes, characters, lighting, and special effects, and uses a high-performance rendering engine to generate high-definition, realistic image sequences.

[0126] III. Sound Generation:

[0127] -Dubbing script to speech conversion: Input the dialogue text of the characters in the script into the speech synthesis model to generate the corresponding voice clips of the characters.

[0128] -Music composition: Based on the script's mood, scene transitions, and other information, generate background music that matches the video's style and rhythm, including theme songs, incidental music, transition music, etc.

[0129] - Sound effects design: Generate various sound effects that match the content of the screen, such as footsteps, door opening and closing sounds, fighting sounds, and environmental noise.

[0130] - Mixing and post-processing: Mixing, balancing, and noise reduction of voice, music, sound effects and other elements to ensure the final audio quality.

[0131] In conclusion, the key to automatically generating video scripts, visuals, and sound through AIGC technology lies in integrating multiple AI algorithms and models, such as NLP, CV, deep learning speech synthesis, and AI composition, to work collaboratively and achieve fully automated creation from text description to audiovisual presentation. These technologies not only significantly improve video production efficiency but also provide creators with unlimited creative space, propelling the video industry into a new era of intelligence.

[0132] Step S130: Obtain the target plot selection control triggered by the current user and the decision score of the target plot selection control.

[0133] Among them, the decision score is a score used to represent the level of safety awareness, which is determined based on the teaching objectives corresponding to the storyline video and the plot development of the target storyline selection control.

[0134] Currently, users can watch videos through dedicated interactive platforms (such as mobile applications, web pages, VR / AR devices, etc.) and interact with them through touch, gestures, voice commands, etc. The platform's built-in feedback system can receive viewer selections, pass them to the AIGC backend for content generation, and synchronize the generated content back to the frontend for display in real time.

[0135] In one example, the algorithm for obtaining the decision score of the triggering target plot selection control includes:

[0136] Using reinforcement learning algorithms, such as Q-Learning or DeepQ-Network (DQN), each user's choice serves as a state transition. The reward function can be designed based on user learning outcomes, satisfaction, and other metrics, enabling the system to continuously optimize its storyline strategy and provide the best personalized experience for the user. Algorithm details:

[0137] The state space `S` contains the user's current contextual state information, the action space `A` includes different story branches that the user can choose, the reward function `R(s,a,s')` measures the user's gain from performing action `a` in state `s` to transition to the new state `s'`, and the policy function `π(a|s)` describes the probability of choosing action `a` in state `s`.

[0138] The goal is to minimize the negative value of the long-term cumulative discount reward:

[0139]

[0140] The state-action value of policy π, or the policy value function, is defined as follows: J(π) is the expected value of the long-term cumulative reward for following policy π. R(s) t a t s t+1 ) indicates that at time step t, an action a is taken. t After from state s t Transition to state s t+1 Instant rewards obtained at that time. This represents the expected calculation when choosing an action according to strategy π. It is the total reward with time discount, which is the sum of the rewards for all future time steps starting from the initial moment, and the reward for each future moment is multiplied by a decay coefficient γ (0≤γ<1), which makes the value of future rewards gradually decrease as they get further away from the current time step.

[0141] Step S140: Based on the plot direction corresponding to the target plot selection control, find the target network security incident cases in the network security incident case set that meet the plot information corresponding to the plot video.

[0142] In practice, the established deep learning natural language processing model is used to analyze the plot direction corresponding to the target plot selection control, the plot information and character information of the previous plot video of the current user, and obtain plot analysis results.

[0143] Based on the results of the plot analysis, different target cybersecurity incident cases that meet the plot information (plot summary and character information) are found in the set of cybersecurity incident cases.

[0144] Specifically, using a bidirectional Transformer model, the probability distribution of the model's output can be obtained through the softmax function. Given an input sequence x and a pre-generated sequence of partial words y1, y2, ..., y... j-1 Predict the word y at position j in the sequence. j The conditional probability can be calculated using the following formula:

[0145] P(y j |y1, y2, ..., y j-1 x) = softmax(Wh j +b)

[0146] Among them, h j The BERT model considers the entire input sequence x and the previously generated words y1, y2, ..., y3. j-1 Then, a hidden state vector is generated for the j-th position. W is a trainable weight matrix used to map the hidden state vector to the predicted score of each word in the vocabulary. b is a trainable bias vector added to the predicted score of each word.

[0147] Furthermore, the target cybersecurity incident case is used as a new current cybersecurity incident case, and the process returns to step S120: Based on the current cybersecurity incident case and the corresponding teaching objectives, AIGC technology is used to generate a narrative video until the current user's viewing time reaches the preset duration.

[0148] Step S150: Based on the correspondence between different decision scores and different security awareness levels in the configuration, and the decision scores corresponding to the plot videos watched by the current user, determine the current user's security awareness level.

[0149] In practice, if the current user's continuous viewing time reaches a preset viewing time threshold, the decision score corresponding to the plot videos watched by the current user is obtained; the decision scores of the watched plot videos are summed to determine the current user's decision score; then, based on the configured correspondence between different decision scores and different security awareness levels, the security awareness level value corresponding to the current user's decision score is determined, and the correspondence between the current user's user identifier and the corresponding security awareness level value is stored.

[0150] To evaluate user decision-making behavior, a binary or multi-class classification problem can be constructed, using the cross-entropy loss function:

[0151]

[0152] Where L represents the average sample loss (Loss), which is typically the objective function to be minimized when training a neural network. N is the number of samples. For each sample i, y i This is the actual label, a binary variable that corresponds to the one-hot encoding of a single class in multi-class classification problems, taking a value of 0 or 1; while in binary classification problems, it directly represents the class label (e.g., 0 represents the negative class, 1 represents the positive class). i This is the probability predicted by the model to belong to the positive class. For multi-class classification problems, after applying the softmax function, for a certain class of the i-th sample, its corresponding p... i This is the predicted probability for that category. log(p) i ) is when y i The log likelihood of the actual class probability when = 1, and log(1-p) i ) is when y i Log likelihood when y = 0 (not the actual class probability). V is the number of classes, y = 0. ij This is the label of sample i corresponding to the true class j, also using one-hot encoding, with only one dimension being 1 and the others being 0. ij This is the probability that the model predicts sample i belongs to class j, given by the softmax function. If we consider the multi-class case of a single sample, the cross-entropy loss between the probability vector output by the softmax function and the corresponding true label ensures that the model learns an effective probability distribution during optimization, making the probability of the correct class as close to 1 as possible, while the probabilities of other classes as close to 0 as possible.

[0153] In some embodiments, AIGC technology can monitor and analyze user decision-making behavior during interaction, providing timely and detailed safety advice and objective evaluation scores, as well as reports upon completion of viewing. It can also dynamically adjust training difficulty and content based on the user's learning progress and performance. Specifically, based on the user's action of selecting controls within the target storyline, behavioral data such as selection preferences, reaction speed, and problem-solving methods are obtained. In-depth analysis of this behavioral data reveals the user's learning characteristics while watching the storyline video, such as learning patterns, strengths and weaknesses, and interest tendencies. These actions may include, but are not limited to, details such as user clicks, dwell time, browsing order, strategies and time spent solving tasks or puzzles, and the number of incorrect attempts. This behavioral data reflects the user's learning habits, comprehension ability, and knowledge gaps, forming the basis for personalized teaching.

[0154] Afterwards, based on learning characteristics, the system provides learning suggestions to the current user. For example, for knowledge points that the user has not yet mastered, the system will recommend corresponding reinforcement exercises or provide more case studies. If the system finds that the user is slow to respond to a certain type of question, it can push targeted tutoring materials or provide clearer guidance.

[0155] In other words, based on user behavior analysis, the personalized content generation module uses AIGC technology to adaptively generate teaching content, prompts, role-playing dialogues, and other elements that match the user's characteristics. It can customize personalized learning paths and interactive scenarios based on factors such as the user's knowledge level, interests, and learning style, ensuring that each viewer receives the most suitable educational experience. For example, it may provide more detailed explanations and examples for beginners, while for users with some prior knowledge, it may focus on advanced threat identification and defense strategy drills.

[0156] Furthermore, when users make choices or perform actions, the system can provide immediate feedback—real-time feedback—indicating whether their decisions are appropriate, explaining the logic behind the correct answers, or providing directions for improvement. The feedback can also adjust its difficulty based on the user's performance, ensuring effective learning at the edge of their comfort zone—neither too easy to cause boredom nor too difficult to cause frustration. Subsequently, as user behavior data accumulates and is analyzed and updated, the system can dynamically adjust the user's learning path to better suit individual needs and developmental stages, achieving a shift from a "one-size-fits-all" approach to a "personalized" educational strategy.

[0157] In some embodiments, online learning algorithms are used to update model parameters as new cybersecurity information emerges, ensuring that the generated video footage always reflects the latest security situation and defense strategies. Algorithm details: Online learning update mechanism: Whenever new cybersecurity threat intelligence or defense strategy data Dnew appears, the model parameters θ will be updated through online gradient descent or other online optimization algorithms.

[0158]

[0159] Where η is the learning rate and L is the loss function, used to measure how well the model fits the new data.

[0160] In some embodiments, as the cybersecurity threat landscape changes and technologies evolve, new threat types can be acquired in real time to identify new cybersecurity incident cases. This updates the cybersecurity incident case set (i.e., the knowledge graph) to reflect the latest threat types, protection measures, and industry trends. The AIGC model can quickly absorb new knowledge input, generate up-to-date video content, and maintain the timeliness of education.

[0161] User experience optimization:

[0162] Interactive video production systems offer the following significant advantages over traditional cybersecurity education materials or videos in terms of enhancing user engagement, improving educational effectiveness, and personalizing learning paths:

[0163] 1. Highly immersive experience:

[0164] Interactive videos, by constructing realistic virtual environments and enabling interactive roles, immerse users in real-world cyber threat scenarios, allowing them to personally experience and respond to various cybersecurity incidents, greatly enhancing the sense of engagement and realism in learning. This immersive experience not only increases user participation but also helps deepen their understanding and retention of cybersecurity issues.

[0165] 2. Immediate feedback and reinforcement learning:

[0166] Users receive immediate feedback from the system on their decisions and actions in the videos. Whether it's positive recognition or negative consequences, everything is presented instantly, creating a highly efficient "learning by doing" teaching model. This instant feedback mechanism helps users quickly correct misconceptions, solidify correct practices, and effectively improve learning efficiency.

[0167] 3. Independent exploration and active learning:

[0168] Interactive videos allow users to drive the plot forward based on their own judgment and choices, and this process of autonomous exploration stimulates users' willingness to learn actively. Users are no longer passively receiving information, but actively participating in the story, solving problems, and avoiding risks, thereby enhancing their initiative and enthusiasm for learning.

[0169] 4. Personalized learning path:

[0170] The system can dynamically adjust the plot, content difficulty, and teaching methods based on user behavior analysis, providing a personalized learning experience. Each user receives content that matches their knowledge level, interests, and learning style, avoiding the mismatch problems caused by a "one-size-fits-all" approach to education and ensuring that the educational content is more targeted and effective.

[0171] 5. Diverse interactions and in-depth participation:

[0172] In addition to storyline selection, the system also supports multimodal interaction, such as voice Q&A, simulated operation, and team collaboration, further enriching the ways users can participate. These diverse interactive methods not only meet the needs of different learning styles but also promote in-depth thinking and practical application, helping to cultivate users' comprehensive cybersecurity skills.

[0173] 6. Continuous updates and adaptability:

[0174] Employing AIGC technology, interactive videos can quickly respond to the latest threat trends and technological developments in the cybersecurity field, updating educational content in real time to maintain its timeliness and relevance. Simultaneously, the system can continuously optimize teaching strategies based on user learning outcome data, ensuring that educational content always maintains optimal educational effectiveness.

[0175] 7. Gamification elements and incentive mechanisms:

[0176] Interactive videos often incorporate gamification elements, such as achievement systems, point rewards, and leaderboards, leveraging the dynamics of competition and cooperation to stimulate users' learning interest and sustained engagement. These incentive mechanisms not only increase the enjoyment of learning but also encourage users to actively practice repeatedly and challenge themselves with more difficult tasks, thereby deepening their learning outcomes.

[0177] In conclusion, interactive video production systems, with their highly immersive, real-time feedback, autonomous exploration, personalized pathways, diverse interactions, continuous updates, and gamified incentives, significantly enhance user engagement, improve the effectiveness of cybersecurity education, and provide a highly personalized learning experience that is difficult to match with traditional educational materials or videos.

[0178] Application scenarios are listed below:

[0179] 1. Corporate employee training:

[0180] Scenario Description: A large multinational corporation adopted an AIGC-based interactive video production system for cybersecurity awareness as part of its annual security training program. The system automatically generates customized interactive video scripts for employees in different departments and positions, covering topics ranging from basic password management and anti-phishing emails to advanced cloud service security and supply chain risk management. Employees log into the company's internal training platform and select video modules relevant to their work for learning.

[0181] 2. School Education:

[0182] Scenario Description: A local education bureau has partnered with us to integrate an interactive video production system into the K-12 information technology curriculum as a crucial component of cybersecurity education. Teachers guide students in watching and participating in interactive videos during class. The videos cover common cybersecurity issues among teenagers, including personal information protection, social network security, and cyberbullying prevention. Furthermore, students can use the school's learning platform for independent review and extended learning after class.

[0183] 3. Public outreach:

[0184] Scenario Description: Cybersecurity regulatory authorities and non-profit organizations jointly launched the "National Cybersecurity Month" campaign. A key initiative was the release of a series of short cybersecurity videos produced using an interactive video production system, which were widely disseminated through social media, video websites, outdoor electronic screens, and other channels. The videos used relatable stories from everyday life, such as online shopping traps, telecom fraud, and the security of smart home devices, guiding viewers to make choices that influenced the plot and ultimately revealing correct security practices.

[0185] 4. Cybersecurity Sand Table Drill

[0186] Scenario Description: Critical information infrastructure operators meticulously crafted a cybersecurity simulation exercise that blended education, practical application, and entertainment. The activity utilized an interactive video-based educational assessment system to construct a highly realistic network environment, encompassing diverse architectures such as enterprise intranets, cloud services, and IoT devices, allowing both attackers and defenders to immerse themselves in responding to various network threats. The defending side directly confronted DDoS attacks, skillfully identifying and handling phishing emails, and accurately detecting and effectively blocking malware intrusions. On-site, realistic visual and sound effects immersed the audience in a tense and exciting simulation, as if they were personally experiencing a real cyber battlefield. The system tracked and recorded the attack and defense dynamics in real time, accurately assessing the cybersecurity attack and defense skills and overall capabilities of both sides, providing immediate feedback and cleverly integrating serious educational assessment into a fun and interactive experience.

[0187] Corresponding to the above method, this application also provides an educational assessment device for interactive videos based on AIGC technology, such as... Figure 2 As shown, the device includes:

[0188] The search unit 210 is used to search for the current network security event case and the corresponding teaching objective in the stored network security event case set based on the current user's user information and the network security type to be evaluated; the network security event case set contains network security event cases of different network security types;

[0189] The generation unit 220 is used to generate a narrative video based on the current network security incident case and the corresponding teaching objectives, using AIGC technology; the video frame at the end of the narrative video includes a narrative selection control that can trigger different narratives;

[0190] The acquisition unit 230 is used to acquire the target plot selection control triggered by the current user and the decision score that triggered the target plot selection control; the decision score is a score used to characterize the level of safety awareness, which is determined based on the teaching objectives corresponding to the plot video and the plot direction of the target plot selection control.

[0191] The search unit 210 is also used to find target network security incident cases in the network security incident case set that satisfy the plot information corresponding to the plot video, based on the plot direction corresponding to the target plot selection control;

[0192] Triggering unit 240 is used to take the target network security incident case as a new current network security incident case and trigger generation unit to execute: based on the current network security incident case and the corresponding teaching objectives, generate a plot video until the current user's viewing time reaches the preset time.

[0193] The determining unit 250 is used to determine the current user's security awareness level based on the correspondence between different configured decision scores and different security awareness level values ​​and the decision scores corresponding to the plot videos watched by the current user.

[0194] The functions of each unit in the interactive video educational assessment device based on AIGC technology provided in the above embodiments of this application can be implemented through the above methods and steps. Therefore, the specific working process and beneficial effects of each unit in the interactive video educational assessment device based on AIGC technology provided in the embodiments of this application will not be repeated here.

[0195] This application also provides an electronic device, such as... Figure 3As shown, it includes a processor 310, a communication interface 320, a memory 330, and a communication bus 340, wherein the processor 310, the communication interface 320, and the memory 330 communicate with each other through the communication bus 340.

[0196] Memory 330 is used to store computer programs;

[0197] When the processor 310 executes the program stored in the memory 330, it performs the following steps:

[0198] Based on the current user's information and the network security type to be evaluated, the system searches the stored network security incident case set for the current network security incident case and the corresponding teaching objective; the network security incident case set contains network security incident cases of different network security types.

[0199] Based on the current cybersecurity incident case and the corresponding teaching objectives, AIGC technology is used to generate a narrative video; the video frame at the end of the narrative video includes a narrative selection control that can trigger different storylines;

[0200] Obtain the target story selection control triggered by the current user and the decision score that triggered the target story selection control; the decision score is a score used to characterize the level of safety awareness, determined based on the teaching objectives corresponding to the story video and the storyline of the target story selection control.

[0201] Based on the plot direction corresponding to the target plot selection control, find the target cybersecurity incident cases in the cybersecurity incident case set that satisfy the plot information corresponding to the plot video;

[0202] The target cybersecurity incident case is used as a new current cybersecurity incident case, and the execution steps are returned: Based on the current cybersecurity incident case and the corresponding teaching objectives, a narrative video is generated until the current user's viewing time reaches the preset duration;

[0203] Based on the correspondence between different configuration decision scores and different security awareness levels, and the decision scores corresponding to the plot videos currently watched by the user, the security awareness level of the current user is determined.

[0204] The communication bus mentioned above can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This communication bus can be divided into address bus, data bus, control bus, etc. For ease of illustration, only one thick line is used to represent it in the diagram, but this does not mean that there is only one bus or one type of bus.

[0205] The communication interface is used for communication between the aforementioned electronic devices and other devices.

[0206] The memory may include random access memory (RAM) or non-volatile memory (NVM), such as at least one disk storage device. Optionally, the memory may also be at least one storage device located remotely from the aforementioned processor.

[0207] The processors mentioned above can be general-purpose processors, including central processing units (CPUs), network processors (NPs), etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0208] The implementation methods and beneficial effects of the various components of the electronic device in the above embodiments for solving the problem can be found in [reference needed]. Figure 1 The steps in the illustrated embodiments are used to implement the electronic device. Therefore, the specific working process and beneficial effects of the electronic device provided in this application will not be repeated here.

[0209] In another embodiment provided in this application, a computer-readable storage medium is also provided, which stores instructions that, when executed on a computer, cause the computer to perform the interactive video educational assessment method based on AIGC technology described in any of the above embodiments.

[0210] In another embodiment provided in this application, a computer program product containing instructions is also provided, which, when run on a computer, causes the computer to execute the interactive video educational assessment method based on AIGC technology described in any of the above embodiments.

[0211] Those skilled in the art will understand that the embodiments in this application can be provided as methods, systems, or computer program products. Therefore, the embodiments in this application can take the form of entirely hardware embodiments, entirely software embodiments, or embodiments combining software and hardware aspects. Furthermore, the embodiments in this application can take the form of computer program products implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0212] This application describes embodiments of methods, apparatus (systems), and computer program products according to embodiments of this application with reference to flowchart illustrations and / or block diagrams. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0213] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0214] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0215] Although preferred embodiments have been described in this application, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the embodiments of this application.

[0216] Obviously, those skilled in the art can make various modifications and variations to the embodiments of this application without departing from the spirit and scope of the embodiments of this application. Therefore, if these modifications and variations to the embodiments of this application fall within the scope of the claims in this application and their equivalents, then this application also intends to include these modifications and variations.

Claims

1. An educational assessment method based on interactive video using AIGC technology, characterized in that, When applied to cybersecurity education systems, the method includes: Based on the current user's information and the network security type to be evaluated, the system searches the stored network security incident case set for the current network security incident case and the corresponding teaching objective; the network security incident case set contains network security incident cases of different network security types. Based on the aforementioned current cybersecurity incident cases and corresponding teaching objectives, AIGC technology is used to generate narrative videos, including: The trained sequence generation model is used to process the network security incident cases and corresponding teaching objectives to obtain plot outlines and character information that meet the teaching objectives. Using a pre-defined AIGC technology, the plot synopsis and character information are combined to generate script text for the cybersecurity incident type; A pre-defined conditional generative adversarial network (GAN) is used to process the script text to obtain a corresponding video clip. This process includes: The generator network in the conditional generative adversarial network generates a sequence of video frames based on the input script text; the video information in each video frame includes scene, character, and action. The discriminator network in the conditional generative adversarial network distinguishes the scenes, characters, and actions in each video frame from the scenes, characters, and actions in the real world. If the discrimination result meets the preset error condition, then the video frame sequence is determined as the plot video corresponding to the script text; The video frame at the end of the story video includes a story selection control that can trigger different storylines; Obtain the target story selection control triggered by the current user and the decision score that triggered the target story selection control; the decision score is a score used to characterize the level of safety awareness, determined based on the teaching objectives corresponding to the story video and the storyline of the target story selection control; Based on the plot direction corresponding to the target plot selection control, find the target cybersecurity incident cases in the cybersecurity incident case set that satisfy the plot information corresponding to the plot video; The target cybersecurity incident case is used as a new current cybersecurity incident case, and the execution steps are returned: Based on the current cybersecurity incident case and the corresponding teaching objectives, a narrative video is generated until the current user's viewing time reaches the preset duration; Based on the correspondence between different configuration decision scores and different security awareness levels, and the decision scores corresponding to the plot videos watched by the current user, the security awareness level of the current user is determined.

2. The method as described in claim 1, characterized in that, The user information includes user identifier and user attribute information; Find any network security incident case and corresponding teaching objective in the stored set of network security incident cases that meet the network security type to be evaluated and the user information, including: Find network security incident cases in the stored network security incident case collection that meet the network security type to be evaluated; If the stored mapping table of user identifiers and security awareness level values ​​contains the user identifier and the corresponding security awareness level value, then select the current network security incident case and the corresponding teaching objective that meet the security awareness level value from the found network security incident cases. If the user identifier is not included in the mapping table between the stored user identifier and the security awareness level value, the initial security awareness level value of the current user is determined according to the user attribute information, and any network security event case that meets the initial security awareness level value and the corresponding teaching objective are selected from the network security event cases found.

3. The method as described in claim 1, characterized in that, Based on the correspondence between different configuration decision scores and different security awareness levels, and the decision scores corresponding to the plot videos currently watched by the user, the security awareness level of the current user is determined, including: If the current user's continuous viewing time reaches a preset viewing time threshold, then obtain the decision score corresponding to the plot video watched by the current user; The current user's decision score is determined by summing the decision scores of the viewed story videos; Based on the correspondence between different configured decision scores and different security awareness levels, the security awareness level corresponding to the current user's decision score is determined. Store the correspondence between the current user's user ID and the corresponding security awareness level value.

4. The method as described in claim 1, characterized in that, Based on the plot development corresponding to the target plot selection control, different target cybersecurity incident cases in the cybersecurity incident case set that satisfy the plot information corresponding to the plot video are found, including: Using the established deep learning natural language processing model, the plot direction corresponding to the target plot selection control, the plot information of the previous plot video of the current user, and the character information are analyzed to obtain plot analysis results; Based on the plot analysis results, different target cybersecurity incident cases that satisfy the plot information are found in the set of cybersecurity incident cases.

5. The method as described in claim 1, characterized in that, The method further includes: Based on the user's action of selecting the target storyline control, obtain the user's behavior data; Deep analysis of the behavioral data yields the learning characteristics of the current user's viewing of the story videos; Based on the learning characteristics, learning suggestions are provided to the current user.

6. An interactive video-based educational assessment device based on AIGC technology, applied in a cybersecurity education system, characterized in that, The device includes: The search unit is used to search for current network security incident cases and corresponding teaching objectives in the stored network security incident case set based on the current user's user information and the network security type to be evaluated; the network security incident case set contains network security incident cases of different network security types; A generation unit is used to generate narrative videos based on the current cybersecurity incident case and corresponding teaching objectives, using AIGC technology; wherein, the generation unit includes: The first processing module is used to process the network security incident case and the corresponding teaching objectives using a trained sequence generation model, and to obtain a plot summary and character information that meet the teaching objectives. The second processing module is used to combine the plot outline and the character information using a preset AIGC technology to generate script text for the network security event type. The third processing module is used to generate an adversarial network using preset conditions to process the script text and obtain the corresponding plot video. Specifically, the third processing module is used for: The generator network in the conditional generative adversarial network generates a sequence of video frames based on the input script text; the video information in each video frame includes scene, character, and action. The discriminator network in the conditional generative adversarial network distinguishes the scenes, characters, and actions in each video frame from the scenes, characters, and actions in the real world. If the discrimination result meets the preset error condition, then the video frame sequence is determined as the plot video corresponding to the script text; The video frame at the end of the story video includes a story selection control that can trigger different storylines; The acquisition unit is used to acquire the target plot selection control triggered by the current user and the decision score that triggered the target plot selection control; the decision score is a score used to characterize the level of safety awareness, which is determined based on the teaching objectives corresponding to the plot video and the plot direction of the target plot selection control. The search unit is also used to find target cybersecurity incident cases in the cybersecurity incident case set that satisfy the plot information corresponding to the plot video, based on the plot direction corresponding to the target plot selection control; The triggering unit is used to take the target network security incident case as a new current network security incident case and trigger the generation unit to execute: based on the current network security incident case and the corresponding teaching objectives, generate a story video until the current user's viewing time reaches the preset time. The determining unit is used to determine the current user's security awareness level based on the correspondence between different configured decision scores and different security awareness levels, and the decision scores corresponding to the plot videos watched by the current user.

7. An electronic device, characterized in that, The electronic device includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; Memory, used to store computer programs; A processor, when executing a program stored in memory, implements the method of any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method described in any one of claims 1-5.

Images