A state machine-based method, equipment, and storage medium for reviewing broadcast control tasks.

Abstract

This application discloses a broadcast control task auditing method, device, and storage medium based on a state machine, belonging to the field of system management technology. The method includes: acquiring a broadcast control task and recording the silent auditing time of the broadcast control task through a state machine; if the silent auditing time is greater than or equal to a preset auditing time threshold, transitioning the state of the broadcast control task from the silent auditing state to the pending execution state through the state machine; after the state machine completes the state transition, generating and storing local audit records of the broadcast control task based on a hash chain; generating a cloud audit chain based on broadcast control instructions issued from the cloud and received execution receipts; cross-validating the hash digest of the local audit record with the hash digest of the cloud audit chain; and executing the corresponding audit action based on the verification result.

Description

Technical Field

[0001] This application relates to the field of system management technology, and in particular to a method, device and storage medium for reviewing broadcast control tasks based on a state machine. Background Technology

[0002] In campus broadcast control task management, manual review is typically used to ensure content security and compliance. After a teacher creates a broadcast control task in the system, the system pushes the task to an administrator account with approval authority. After logging in, the administrator reviews the task content one by one, and clicks "approval" after confirming that everything is correct before the task enters the execution queue.

[0003] Because each task requires manual review, when there are many tasks or when reviewers are temporarily unavailable, tasks may miss their scheduled execution time due to delayed approval. Therefore, if a task has not received manual approval before the scheduled time, the broadcast control system usually approves it automatically. However, in this situation, the system cannot distinguish the risk level of the task content. For tasks involving sensitive content, automatic approval due to time constraints could lead to the broadcast of incorrect or inappropriate information, causing teaching accidents or management disputes.

[0004] The above content is only used to help understand the technical solution of this application and does not represent an admission that the above content is prior art. Summary of the Invention

[0005] The main purpose of this application is to provide a broadcast control task review method, device and storage medium based on state machine, which aims to solve the technical problem of insufficient accuracy of broadcast control system in reviewing sensitive content.

[0006] To achieve the above objectives, this application provides a state machine-based method for reviewing broadcast control tasks, the method comprising the following steps: Obtain the broadcast control task and record the silent review time of the broadcast control task through a state machine; If the silent review time is greater than or equal to the preset review time threshold, the state of the broadcast control task is transitioned from the silent review state to the pending execution state through the state machine. After the state machine completes the state transition, local audit records for the broadcast control task are generated and stored based on the hash chain. Based on the broadcast control instructions issued from the cloud and the received execution receipts, a cloud audit chain is generated; The hash digest of the local audit record is cross-validated with the hash digest of the cloud audit chain, and the corresponding audit action is executed based on the verification result.

[0007] In one embodiment, before the step of acquiring the broadcast control task and setting the state of the broadcast control task to a silent review state via a state machine, the method further includes: Determine whether the content template for the broadcast control task is being used for the first time; If so, set the status of the broadcast control task to pending confirmation. Upon receiving the confirmation instruction for the broadcast control task, the content template is marked as a confirmed template, and the status of the broadcast control task is set to the silent review status.

[0008] In one embodiment, before the step of transitioning the state of the broadcast control task from the silent review state to the pending execution state via the state machine when the silent review time is greater than or equal to a preset review time threshold, the method further includes: The content of the broadcast control task is risk-scored, and the risk level of the broadcast control task is marked as Level 1, Level 2 and / or Level 3 based on the risk score. If the risk level is Level 2, then the timer for the silent review period is paused, and the status of the broadcast control task is moved to the pending confirmation status. If the risk level is Level 3, then stop the timer and change the status of the broadcast control task to the blocked state.

[0009] In one embodiment, the step of performing a risk score on the content of the broadcast control task and marking the risk level of the broadcast control task as a first level, a second level, and / or a third level based on the risk score includes: Obtain the characteristic parameters of the current scene, including the current time period type, the historical interception rate of the user executing the broadcast control task, and the sensitivity level of the terminal involved in the broadcast control task; The risk score and the feature parameters are input into the reinforcement learning model to obtain the risk threshold boundaries of the first, second and third levels output by the reinforcement learning model. The risk level of the broadcast control task is determined based on the comparison between the risk score and the risk threshold boundary. After the broadcast control task is completed, a reward value is calculated based on the user's feedback on the broadcast control task, and the parameters of the reinforcement learning model are updated based on the reward value.

[0010] In one embodiment, before the step of transitioning the state of the broadcast control task from the silent review state to the pending execution state via the state machine when the silent review time is greater than or equal to a preset review time threshold, the method further includes: Obtain the confirmation status identifier, risk level identifier, timer expiration identifier, and interception event identifier of the content template of the broadcast control task; The confirmation status identifier, the risk level identifier, the timer expiration identifier, and the interception event identifier are concatenated in a preset order to form a first determination vector; Using the first decision vector as input, the transition conditions of the state machine are determined; Based on the determination result, the first determination vector is combined with the current state identifier of the state machine to form a second determination vector. A hash operation is performed on the second determination vector to generate a migration certificate. The migration certificate is written into the local audit record and then the state migration is performed.

[0011] In one embodiment, the state machine-based broadcast control task review method further includes: Receive modification or interception instructions for the broadcast control task; If a modification instruction is received, the state machine is used to transition the state of the broadcast control task to the draft state, the silent review time is cleared, and the state of the broadcast control task is set to the silent review state after the broadcast control task is edited. If an interception command is received, the state machine is used to transition the state of the broadcast control task to the intercepted state, and the timer is stopped. If an emergency release instruction is received, the status of the broadcast control task will be moved to the pending execution status.

[0012] In one embodiment, after the steps of acquiring the broadcast control task and recording the silent review time of the broadcast control task through a state machine, the method further includes: Obtain the content fingerprint of the content payload in the broadcast control task; The content fingerprint is matched with each abnormal fingerprint in the abnormal content fingerprint database based on similarity. If the similarity obtained by matching is greater than or equal to the preset similarity threshold, the state of the broadcast control task is transferred to the blocked state through the state machine, and the silent review time is cleared.

[0013] In one embodiment, the state machine-based broadcast control task review method further includes: Analyze the temporal dependencies between the broadcast control task and other broadcast control tasks to generate a list of dependent tasks; When there are tasks in the dependent task list of the broadcast control task that are not marked as completed or approved, the status of the broadcast control task is set to waiting for dependency, and the silent review time is cleared. Once all tasks in the dependent task list have a completed or approved status, the broadcast control task will be restored to the silent review status.

[0014] In addition, to achieve the above objectives, this application also provides a state machine-based broadcast control task review device, the device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the state machine-based broadcast control task review method as described above.

[0015] In addition, to achieve the above objectives, this application also provides a storage medium, which is a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the state machine-based broadcast control task review method as described above.

[0016] One or more technical solutions proposed in this application have at least the following technical effects: This application uses a state machine to set the task's state to a silent review state and starts a timer. It then waits for the timer to expire. When the timer expires and no interception instruction for the broadcast control task is received, the task state is transitioned from the silent review state to the pending execution state. Simultaneously, a local audit record is generated and stored based on a hash chain during each state transition. A cloud audit chain is also generated independently of the local audit record. The hash digest of the local audit record is cross-validated with the hash digest of the cloud audit chain. If the validation fails, an audit alarm is triggered. This dual-track independent recording and cross-validation mechanism ensures both the efficiency of the automated review process and the traceability of the state transition process and the immutability of the audit data. This allows automatically approved tasks to be reliably reviewed afterward, effectively solving the security and liability determination risks inherent in simple timeout automatic approval mechanisms. Attached Figure Description

[0017] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating the first embodiment of the broadcast control task review method based on a state machine in this application; Figure 2 This is a flowchart illustrating the second embodiment of the broadcast control task review method based on a state machine according to this application. Figure 3 This is a flowchart illustrating the third embodiment of the broadcast control task review method based on a state machine in this application; Figure 4 This is a schematic diagram of the structure of a state machine-based broadcast control task review device in the hardware operating environment involved in the embodiments of this application.

[0020] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0021] It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to limit this application.

[0022] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0023] The main solution of this application embodiment is as follows: obtain the broadcast control task, and record the silent review time of the broadcast control task through a state machine. If the silent review time is greater than or equal to a preset review time threshold, the state of the broadcast control task is transitioned from the silent review state to the pending execution state through the state machine. After the state machine completes the state transition, the local audit record of the broadcast control task is generated and stored based on the hash chain. According to the broadcast control instructions issued by the cloud and the received execution receipt, a cloud audit chain is generated. The hash digest of the local audit record is cross-validated with the hash digest of the cloud audit chain, and the corresponding audit action is executed according to the verification result.

[0024] Current technology in campus broadcast control task management typically relies on manual review to ensure content security and compliance. After a teacher creates a broadcast control task in the system, the system pushes the task to an administrator account with approval authority. The administrator logs in, reviews each task's content, and clicks "approval" after confirming its accuracy before the task enters the execution queue. Because each task requires manual review, when there are many tasks or the reviewer is temporarily unavailable, tasks may miss their scheduled execution time due to delayed approval. Therefore, if a task has not received manual approval before the scheduled time, the broadcast control system usually approves it automatically. However, in this case, the system cannot distinguish the risk level of the task content. For tasks involving sensitive content, automatic approval due to time constraints could lead to the broadcast of incorrect or inappropriate information, causing teaching accidents or management liability disputes.

[0025] This application uses a state machine to set the task's state to a silent review state and starts a timer. It then waits for the timer to expire. When the timer expires and no interception instruction for the broadcast control task is received, the task state is transitioned from the silent review state to the pending execution state. Simultaneously, a local audit record is generated and stored based on a hash chain during each state transition. A cloud audit chain is also generated independently of the local audit record. The hash digest of the local audit record is cross-validated with the hash digest of the cloud audit chain. If the validation fails, an audit alarm is triggered. This dual-track independent recording and cross-validation mechanism ensures both the efficiency of the automated review process and the traceability of the state transition process and the immutability of the audit data. This allows automatically approved tasks to be reliably reviewed afterward, effectively solving the security and liability determination risks inherent in simple timeout automatic approval mechanisms.

[0026] To better understand the above technical solutions, exemplary embodiments of this application will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of this application are shown in the drawings, it should be understood that this application can be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of this application and to fully convey the scope of this application to those skilled in the art.

[0027] It should be noted that the executing entity in this embodiment can be a broadcast control task review system, or a computing service device with data processing, network communication, and program execution functions, such as a tablet computer, personal computer, or mobile phone, or an electronic device capable of implementing the above functions, a state machine-based broadcast control task review device, etc. This embodiment does not specifically limit it. The following uses a broadcast control task review system as an example to describe this embodiment and the following embodiments.

[0028] Based on this, embodiments of this application provide a state machine-based method for reviewing broadcast control tasks, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the broadcast control task review method based on a state machine according to this application.

[0029] In this embodiment, the state machine-based broadcast control task review method includes steps S10 to S40: Step S10: Obtain the broadcast control task and record the silent review time of the broadcast control task through the state machine; In this embodiment, the broadcast control system continuously monitors task creation events or periodically polls the task database. When a new broadcast control task is created or activated, the system reads the complete data record of the task from the database. The system calls the state machine engine to perform a state initialization operation on the task, setting the task's state field to silent review state. Simultaneously, it obtains the current timestamp from the system clock and records this time as the start time of the silent review in the task object's time attribute field. The timer module bound to the task is started, running continuously throughout the task's silent review period, periodically calculating and updating the cumulative silent review time value. The state machine predefines multiple possible states for the broadcast control task, such as draft, scheduled, silent review, pending confirmation, intercepted, pending execution, being distributed, running, completed, failed, and rolled back states. It also predefines legal transition paths between these states, performing state transitions or state changes through the state machine engine's transition interface. The silent review time refers to the cumulative length of time the broadcast control task has been in the silent review state.

[0030] Specifically, the system performs precise retrieval based on task identifiers, receiving new task requests submitted through the task management interface or broadcast control task data packets pushed by the upstream orchestration system via the application programming interface (API). For new tasks submitted through the interface, the system parses the submitted tabular data, extracting the task name, planned execution time, target region identifier, content payload Uniform Resource Locator (URL), and priority fields. Based on a predefined data model, these fields are assembled into structured task objects and written to the task database. For task data packets pushed via the API, the system directly parses the structured data carried in the data packet, converts it into internal task objects through field mapping relationships, and also writes it to the task database. After successful task data writing, the system calls the initialization interface of the state machine engine. This interface allocates a state slot for the task in the state machine's internal state table, sets the current state value of the slot to silent review, and writes the system's current timestamp into the start time field of the slot. The timer module is implemented using an asynchronous timer. The system creates an independent timer object for each task in the silent review state. The timer is triggered at fixed time intervals. When triggered, the difference between the current system time and the start timestamp is calculated, and the calculation result is stored in the silent review time field of the task.

[0031] As an optional implementation, the system first checks the initial usage status of the content templates used by the broadcast control task before initiating the silent review state. The system maintains a template usage record table in the task database, recording the task identifier and usage time of each template identifier's first use. When the system receives a new broadcast control task, it extracts the template identifier of the content template referenced by the task and queries the template usage record table using this template identifier as the key. If the query result is empty, it means that the template has not been used by any task. The system does not call the state machine engine to set the task to a silent review state, but instead sets the task status to a pending confirmation state and sends confirmation request notifications to both the task creator and the administrator account. When the system receives a confirmation instruction from the administrator through the management interface, it writes the record corresponding to the template identifier contained in the confirmation instruction into the template usage record table, simultaneously transitioning the task status from the pending confirmation state to the silent review state and performing a start timestamp recording operation. If the query result is not empty, it means that the template has been confirmed in a previous task, and the system directly enters the silent review state.

[0032] As an alternative implementation, after initializing the silent review state, the system performs a risk score on the content of the broadcast control task. The system inputs the text portion of the task's content payload into a risk scoring model, which is pre-trained using sample corpora labeled with risk levels, and outputs a continuous risk score. The system compares this continuous risk score with multiple preset threshold boundaries, which divide the risk score range into three levels: Level 1 (lowest risk), Level 2 (medium risk), and Level 3 (highest risk). If the risk level is Level 2, the system pauses the timing operation of the timer module bound to the task, transitions the task status to the pending confirmation state, and pushes a risk warning message to the administrator. Only after the administrator confirms safety does the system resume the timer and return to the silent review state. If the risk level is Level 3, the system stops the timer module and directly transitions the task status to the blocked state, permanently excluding the task from the automatic review process.

[0033] In a specific implementation scenario, a teacher creates a broadcast control task for a school-wide fire drill notification through the broadcast control system. After task creation, the system calls the state machine engine to initialize the task's state to silent review, and records the current time as the start time of the silent review. The system then starts an accumulating timer. Each time the timer is triggered, the system reads the current system time and subtracts the start time, updating the silent review time field of the task with the difference. If the content template of the task is being used for the first time, the system sets the task status to pending confirmation and sends a confirmation request to the administrator. Only after the administrator confirms does the task enter the silent review state and start timing. If the system's risk scoring model outputs a risk score for the task's content that falls within the third-level risk range, the system directly transitions the task status to the blocked state, and the notification will not automatically pass the review.

[0034] Optionally, after acquiring a broadcast control task, before starting the silent review timer, or during the timer's operation, the broadcast control system performs intelligent security detection on the task's content payload. It acquires the content fingerprint of the content payload in the broadcast control task, matches this fingerprint with various abnormal fingerprints in the abnormal content fingerprint database, and if the similarity score is greater than or equal to a preset similarity threshold, it transitions the broadcast control task's state to the blocked state via a state machine and clears the silent review time.

[0035] Specifically, the system executes different content fingerprint extraction strategies based on the data type of the content payload. When the content payload is plain text, the system uses the term frequency inverse document frequency (TNF) vectorization method, which calculates the TNF value of each word after text segmentation and combines them to form a text feature vector as the content fingerprint. When the content payload is an image, the system uses a perceptual hashing algorithm, which calculates the discrete cosine transform coefficients after scaling the image to a fixed size, extracts low-frequency coefficients to generate a fixed-length binary hash value as the content fingerprint. When the content payload is video, the system extracts the keyframe sequence of the video, calculates the perceptual hash value for each keyframe, and averages the hash values ​​of all keyframes as the video's content fingerprint. The system calculates the similarity between the extracted content fingerprint and each abnormal fingerprint in the abnormal content fingerprint database. For text feature vectors, the system calculates the cosine similarity between two vectors. For perceptual hash values, the system calculates the Hamming distance between two hash values ​​and normalizes it into a similarity score. The system selects the maximum value from all similarity scores and compares this maximum value with a preset similarity threshold. When the maximum value is greater than or equal to the preset similarity threshold, the system determines that the match is successful, immediately calls the state machine engine to transition the task state from the current state to the intercepted state, and simultaneously calls the timer module's clear interface to reset the silent review time to zero. The task is then removed from the automatic review process. When the maximum value is less than the preset similarity threshold, the system determines that the match has failed, and the task maintains its current state and continues to execute the normal process.

[0036] Step S20: If the silent review time is greater than or equal to the preset review time threshold, the state of the broadcast control task is transitioned from the silent review state to the pending execution state through the state machine; In this embodiment, the broadcast control system continuously monitors the cumulative silent review time of all broadcast control tasks in a silent review state at a preset polling cycle. In each poll, the system traverses the list of tasks in the silent review state, reads the current cumulative silent review time of each task in the list, and compares this value with a preset review time threshold. When the cumulative silent review time of a task reaches or exceeds the preset review time threshold, the system determines that the task has met the time conditions for automatic silent review approval, constructs a state transition request, and submits it to the state machine engine. Upon receiving the transition request, the state machine engine performs a validity check, confirming that the current task state is silent review and the target state is pending execution. The engine then performs a state update operation, changing the task's state field from silent review to pending execution, and records the timestamp and triggering conditions of the transition. The pending execution state is an intermediate state in the state machine indicating that the broadcast control task has passed review and is waiting to reach the planned execution time. The preset review time threshold is a pre-configured time length value, representing the maximum duration for which a broadcast control task is allowed to wait in the silent review state.

[0037] Specifically, the system executes monitoring and comparison logic through a background scheduling thread, which repeats the process at fixed time intervals. During each execution, the scheduling thread constructs a database query, with the query conditions set to the status field equal to the silent review status and the planned execution time later than the current system time. The system retrieves all tasks that meet the conditions from the task database and calculates the current silent review time for each task. The system determines the applicable review time threshold for each task according to a preset threshold configuration table, which configures differentiated thresholds for different task types, target regions, or time periods. The system compares the current silent review time of the task with the applicable threshold. If the current silent review time is less than the threshold, the system does nothing, and the task continues to wait for the next poll. If the current silent review time is greater than or equal to the threshold, the system encapsulates the task identifier and current task status information into a state transition request and calls the state machine engine's transition interface. The state machine engine internally maintains a state transition matrix, where the row index represents the current state, the column index represents the target state, and each cell in the matrix stores a boolean value indicating whether the transition path is valid. Upon receiving a migration request, the engine accesses the migration matrix with silent review status as rows and pending execution status as columns. If the corresponding boolean value is true, the engine executes the migration operation and writes the migration record to the state migration log table. To ensure the atomicity of state migration and data consistency, the state machine engine starts a database transaction when executing the migration operation. Within the transaction, both state field updates and migration log insertion operations are performed simultaneously. When the transaction is committed, it ensures that both operations either succeed or are rolled back.

[0038] As an optional implementation, the system additionally performs a four-condition joint judgment process before executing state transition. The system obtains the current values ​​of four flags: the content template confirmation status flag, the risk level flag, the timer expiration flag, and the interception event flag. The template confirmation status flag indicates whether the content template used by the task has been manually confirmed; the risk level flag indicates whether the task's risk level is Level 1; the timer expiration flag indicates whether the silent review period has reached the threshold; and the interception event flag indicates whether any interception instructions were received during the task's silent review period. The system concatenates these four flags in a preset order to form a first judgment vector. The state machine engine's transition condition judgment module receives this first judgment vector as input and matches it against a preset judgment rule table. Transition is only allowed when all four flags are true. After successful judgment, the system concatenates the first judgment vector with the current state flag of the state machine to form a second judgment vector. A secure hash algorithm is then applied to this string to generate a fixed-length hash value as the transition credential for this operation. This transition credential is subsequently written into the audit log.

[0039] As an alternative implementation, the system continuously receives operation instructions for broadcast control tasks during the polling wait period. When the system receives a modification instruction, it does not wait for the timer to expire, but immediately transitions the task status from silent review to draft status via the state machine and clears the accumulated silent review time value for the task. Simultaneously, the system pauses the timer module. After the task enters the draft status, the task creator or administrator can edit the task content, adjust the planned execution time, or change the target terminal range. After editing, the user resubmits the task, and the system resets the task status to silent review and resets the start timestamp to restart the timing. When the system receives an interception instruction, it immediately transitions the task status to the intercepted state via the state machine and destroys the timer module. When the system receives an emergency release instruction, it skips all condition checks and directly transitions the task status from the current state to the pending execution state. Emergency content can enter the distribution process immediately at any time without waiting for the silent review time threshold to be reached.

[0040] Optionally, before the broadcast control system performs the transition operation from the silent review state to the pending execution state, it needs to execute a multi-condition joint judgment process. This involves obtaining the confirmation status identifier, risk level identifier, timer expiration identifier, and interception event identifier of the content template of the broadcast control task. These identifiers are then concatenated in a preset order to form a first judgment vector. This first judgment vector serves as input to determine the transition conditions of the state machine. Based on the judgment result, the first judgment vector is combined with the current state identifier of the state machine to form a second judgment vector. A hash operation is performed on the second judgment vector to generate a transition certificate. After writing the transition certificate to the local audit record, the state transition is executed.

[0041] After determining that the silent review time has reached the threshold, the system first calls the status query interface to obtain the current values ​​of four flags. The confirmation status flag indicates whether the content template used by the task has been manually confirmed; the timer expiration flag indicates whether the silent review time has reached the preset review time threshold; this flag is set to true by the timer module when it detects that the silent review time has reached the threshold; and the interception event flag indicates whether any interception commands were received during the task's silent review state. The system concatenates the four Boolean values ​​into a four-bit binary string in the order of confirmation status flag, risk level flag, timer expiration flag, and interception event flag, forming the first judgment vector. The state machine engine's transition condition judgment module internally maintains a judgment rule table. This table uses the first judgment vector as an index and Boolean values ​​as output values. Only the output corresponding to "1111" is true; all other combinations are false. The module queries this table using the first judgment vector as the key to obtain the judgment result. If the judgment result is true, the system concatenates the first judgment vector with the current status flag value of the state machine into a string to form the second judgment vector. A secure hash algorithm is then used to operate on this string to generate a fixed-length hash value as a transition credential. The system writes the migration credentials as an additional field into the local audit log, and then calls the state machine engine to perform the state migration operation.

[0042] Optionally, the broadcast control system continuously listens for operation commands from the user interface or application programming interface while the task is in silent review mode and the timer is running. The system receives three types of commands: modification commands, interception commands, and emergency release commands. Each command triggers different state transition paths and timer control operations. Specifically, the system receives operation commands through a message queue or Hypertext Transfer Protocol interface. Each command includes a task identifier, command type, and operator identity fields. After parsing the command, the system queries the current task's status based on the task identifier. When the command type is a modification command, the system first verifies whether the operator has editing permissions for the task. If the permission verification is successful, the system calls the state machine engine's transition interface to transition the task status from silent review mode to draft mode. Simultaneously, the system calls the timer module's clear interface to reset the accumulated silent review time to zero. After the task enters the draft mode, the system returns the editing interface to the user, allowing the user to modify the task content, adjust the planned execution time, or change the target terminal range. After the user completes editing and resubmits, the system writes the updated task data to the database, then calls the state machine engine to reset the task status from draft to silent review, and re-records the start timestamp and starts the timer. When the instruction type is an intercept instruction, the system calls the state machine engine to transition the task status from silent review to intercepted, and simultaneously calls the timer module's destruction interface to stop the timer and release timer resources. Tasks in the intercepted state will not trigger any automatic approval migration. When the instruction type is an emergency release instruction, the system first verifies whether the operator has emergency release permissions. After the permission verification is successful, the system skips all condition checks and timer checks, and directly calls the state machine engine to transition the task status from the current state to the pending execution state. The task immediately enters the dispatch scheduling queue to wait for execution.

[0043] Step S30: After the state machine completes the state transition, generate and store the local audit records of the broadcast control task based on the hash chain; In this embodiment, the state machine engine triggers an audit log generation event after successfully completing a state transition operation. The audit log generation module of the broadcast control system listens for this event and is activated when the event is triggered. It reads relevant information about this transition from the transition context of the state machine engine, including the task identifier, the state before the transition, the state after the transition, the transition timestamp, and the condition data that triggered this transition. The audit log generation module concatenates this information with the hash value of the previous audit log stored in the system, performs a one-way hash function operation on the concatenation result to obtain the hash value of the current audit log, and then writes the structured data of the current audit log along with the calculated hash value into the local audit database. At the same time, it saves the current hash value for use as the preceding hash value when generating the next audit log. A hash chain is a cryptographic data structure in which each data block contains the hash value of the previous data block. When the content of any data block in the chain is modified, the hash value of that data block changes, causing the preceding hash values ​​stored in all subsequent data blocks to not match the recalculated hash values. Local audit logs are structured data entries stored on the local storage medium of the broadcast control system, used to record key event information at each stage of the broadcast control task's lifecycle.

[0044] Specifically, after activation, the audit log generation module first queries the local audit database to determine if the specified task already has historical audit logs. The module executes a counting query to obtain the number of existing audit logs for the task. If the count is zero, the module uses a preset initial hash value as the preceding hash value. If the count is greater than zero, the module queries the latest audit log in descending order of task identifier and audit log number, and reads the hash value of the previous record from the hash value field of that record as the preceding hash value for this generation. The module then assembles the data payload of the current audit log, which includes a list of key-value pairs containing the task identifier, pre-migration state name, post-migration state name, migration timestamp, and trigger conditions. The module serializes the data payload into a string format, concatenates this string with the preceding hash value using a specific delimiter to form a complete input string, and performs a cryptographically secure hash algorithm on this input string to obtain a fixed-length output hash value. The module inserts the data payload of the current audit log, the preceding hash value, the calculated current hash value, and the record generation timestamp as a complete data entry into the local audit database. After the insertion operation is completed, the module updates the current hash value in the task context for use when generating subsequent audit logs.

[0045] As an optional implementation, the audit log generation module additionally writes the migration credential generated during the state transition process into the current audit log when generating local audit logs. The migration credential is the hash value of the second decision vector, which contains a combination of four flags: the task's confirmation status identifier, risk level identifier, timer expiration identifier, and interception event identifier at the time of migration. After writing the migration credential into the audit log, the audit log records the static information of the migration event and the complete conditional state at the time the migration event occurred. Post-event verification involves rereading the task's conditional state at the time of migration and recalculating the hash value. The calculated result is compared with the migration credential stored in the log to confirm whether the four conditions were indeed simultaneously met when the migration occurred.

[0046] As an alternative implementation, the system first parses the timing dependencies between the current broadcast control task and other broadcast control tasks before generating local audit logs. The system extracts a list of upstream task identifiers referenced by the task content or reads task dependency definitions from the task orchestration configuration to generate a dependent task list. If the dependent task list is not empty, the system queries the task status of each dependent task. When at least one dependent task is found to be in a state other than completed or approved, the system does not perform a state transition, but instead sets the current task's status to a waiting-for-dependency state and clears the accumulated silent review time. Whenever a dependent task's status changes, the system re-checks the status of all dependent tasks. When all tasks in the dependent task list are in a completed or approved state, the system restores the current task's status from the waiting-for-dependency state to the silent review state and restarts the timestamp for timing.

[0047] Optionally, before receiving a broadcast control task and determining that the silent review time has reached a threshold, the broadcast control system also needs to parse the temporal dependencies between the current broadcast control task and other broadcast control tasks. Parsing the temporal dependencies between the broadcast control task and other broadcast control tasks generates a list of dependent tasks. When there are tasks in the list of dependent tasks that are not marked as completed or approved, the broadcast control task's status is set to a waiting-for-dependency state, and the silent review time is cleared. Once all tasks in the list of dependent tasks are marked as completed or approved, the broadcast control task's status is restored to the silent review state.

[0048] Specifically, the system parses the task's data payload during task creation or content update, identifying other task identifiers referenced within it. Dependencies between tasks are defined by the upstream system during task orchestration and stored in the task data as structured fields. Each field contains an array of task identifiers, where each element represents an upstream task that the current task depends on. The system reads this array to generate a list of dependent tasks. When the silent review time reaches a threshold, the system first checks if the task contains a non-empty list of dependent tasks before triggering a state transition. If the list is empty, the system proceeds with the normal transition from the silent review state to the pending state. If the list is not empty, the system iterates through each task identifier in the list, querying the task status database for each identifier to obtain the current status value of each dependent task. The system then checks each status value to see if it belongs to the completed or approved status set. When any status value does not belong to this set, the system stops iterating, determines that the dependency condition is not met, calls the state machine engine to transition the current task state to the waiting-for-dependencies state, and simultaneously calls the timer module's clear interface to reset the silent review time to zero. The system registers listener callbacks for dependency task status change events. Whenever the status of a dependency task changes, the callback function is triggered and dependency condition checks are re-executed. When the status of all dependency tasks belongs to the "completed" or "approved" set, the system calls the state machine engine to restore the current task status from the "waiting for dependencies" state to the "silent review" state, re-records the start timestamp, starts the timer, and causes the task to re-enter the silent review time accumulation process.

[0049] Step S40: Generate a cloud audit chain based on the broadcast control instructions issued by the cloud and the received execution receipt; In this embodiment, while issuing broadcast control commands to edge gateways or terminal devices, the cloud system generates a cloud-side audit record based on the command and writes it to the cloud audit database. The data source for the cloud audit record is the broadcast control command data packet issued by the cloud system itself and the execution receipt data packet subsequently received from the edge gateway or terminal device. Each time a broadcast control command is issued, the cloud system extracts the task identifier, target terminal identifier, command content summary, and issuance timestamp from the command and writes them into the cloud audit database as the basic data for the cloud audit record. It also uses a hash chain method to embed the hash value of the previous cloud audit record into the current record, forming an independent hash audit chain on the cloud side. When the cloud system receives an execution receipt from the edge gateway, it associates the task identifier and command identifier in the receipt with the previously written cloud audit record and supplements the record with the execution result status, actual execution timestamp, and error code. The cloud audit chain is independently stored in the audit record sequence on the cloud server, generated independently based on the broadcast control commands issued by the cloud system itself and the execution receipts received from the terminal or edge gateway, forming a dual-track parallel audit data system with the local hash chain audit records.

[0050] Specifically, when executing broadcast control command issuance operations, the cloud system first constructs a cloud audit record object. The initialization of this object begins during the command construction phase. The system extracts the task identifier and planned execution time from the broadcast control task object, the terminal identifier and terminal type from the target terminal configuration, and calculates the hash value of the content summary from the content payload. These fields are then filled into the corresponding attributes of the cloud audit record object. Simultaneously, the system reads the hash value of the previous cloud audit record for this task from the cloud audit database as the preceding hash value. If there is no cloud audit record for this task, the initial hash value of the cloud system is used. After serializing the filled cloud audit record object into a storage format, the system concatenates it with the preceding hash value and calculates the hash value. The record data, the preceding hash value, and the current hash value are then written to the cloud audit database, completing the generation of the audit record at the time of issuance. During writing, the execution status field of this record is marked as issued and awaiting receipt. When the cloud system receives an execution receipt from the edge gateway via a message queue or transmission control protocol connection, the system parses the receipt data packet and extracts the task identifier, execution status code, actual execution timestamp, and error description fields. The system queries the cloud audit database based on the task identifier, finds the corresponding cloud audit record that has been issued and is awaiting receipt, writes the execution status code into the execution result field, writes the actual execution timestamp into the receipt timestamp field, writes the error description into the error information field, and then updates the status of the record to "receipt received".

[0051] As an optional implementation, if a broadcast control task is transitioned to an blocked state during the review stage due to an abnormal content fingerprint matching, the task will not generate any cloud-based broadcast control instructions, and there will be no corresponding broadcast control instruction issuance record in the cloud audit chain. When the system performs subsequent cross-validation, it detects a transition record between the silent review state and the blocked state of the task in the local audit records, but there is no corresponding instruction issuance record or receipt record in the cloud audit chain. In this case, the system recognizes this state as a normal difference under intelligent blocking scenarios and does not judge it as data missing or tampered with. The system then marks the task as having entered the blocked audit closed-loop state in the audit report.

[0052] As an alternative implementation, the cloud system generates a separate cloud audit sub-chain for each broadcast control task when generating the cloud audit chain, instead of mixing all tasks into a single global audit chain. Each task's cloud audit sub-chain independently maintains its hash chain structure. The head hash value of the sub-chain is generated when the task is created, and the tail hash value is updated each time a new record is appended to the sub-chain. The system periodically aggregates the tail hash values ​​of the cloud audit sub-chains of all active tasks and calculates a global verification hash value, which is then uploaded to a third-party evidence storage platform.

[0053] Step S50: Cross-validate the hash digest of the local audit record with the hash digest of the cloud audit chain, and execute the corresponding audit action based on the verification result.

[0054] In this embodiment, the cloud system or a dedicated audit verification service periodically reads the hash digests of local audit records and cloud audit chains from the broadcast control task, comparing the two to verify the consistency of the audit data. The verification process first compares whether the end hash values ​​of the two audit chains are the same; if they are the same, the audit data is considered complete and consistent. If the end hash values ​​are different, the system further backtracks to each corresponding record in the two audit chains, comparing the key field values ​​of each record entry to determine the specific location and content of the difference. The system generates corresponding audit actions based on the nature and severity of the difference, including recording a difference log, generating an alarm notification, marking abnormal tasks, preventing further distribution, or triggering manual review. Cross-validation determines data consistency and completeness by comparing audit information from different data sources, utilizing the characteristics of hash chains to ensure that the two audit chains should ideally be completely consistent. A hash digest is a fixed-length hash value obtained by operating a one-way hash function on the complete content of an audit record, uniquely representing the content of the original audit record.

[0055] Specifically, the audit verification service is carried out by an independent module in the cloud system, which does not depend on any other module in the main broadcast control process. The verification service performs verification operations at the task-by-task level, receiving the task identifier as an input parameter. The verification service connects to both the local and cloud audit databases, querying all audit records for that task from each database based on the task identifier. The verification service extracts the hash value of the last record from the local audit records as a local hash digest, and extracts the hash value of the last record from the cloud audit chain as a cloud hash digest. It first compares whether the two hash digests are completely identical. If they are identical, the verification service generates a verification pass result, records the verification timestamp, and ends the verification process. If they are inconsistent, the verification service enters the difference localization phase, traversing backwards from the first record in both audit chains. For each record index position, it calculates and compares the hash values ​​of the local and cloud audit records at that position. When it reaches the first position where the hash values ​​differ, the verification service identifies that position as the starting point of the difference, extracts all field values ​​from the two records at that position, and compares them field by field to accurately locate the differing fields. The verification service encapsulates the starting point of the difference, the name of the difference field, and the content of the difference into a difference report, and performs corresponding audit actions based on the content of the difference report.

[0056] As an optional implementation, after comparing the end hash values ​​of the two audit chains, even if the end hash values ​​are the same, the verification service still randomly selects several non-end records from the two audit chains for hash comparison. The sampling strategy uses time-stratified random sampling, drawing sample records from evenly distributed positions on the timeline of the audit chains. For the selected record indices, the hash values ​​of the local audit record and the cloud audit record are calculated and compared. This method can detect potential internal data inconsistencies within the entire audit chain even if the end hash values ​​are the same. If the random sampling verification finds a difference, subsequent processing is performed according to the same difference location and audit actions.

[0057] As an alternative implementation, during cross-validation, if a broadcast control task experiences a waiting-dependency state during execution due to dependency management, the validation service performs specific verification on the dependency-related audit records. The validation service identifies the time points from the audit records when a task enters and recovers from the waiting-dependency state, and extracts the status change records of each task in the dependency task list within the time period between these two time points. The validation service compares the dependency task status change sequence recorded in the local audit records with the corresponding task status change sequence recorded in the cloud audit chain to confirm whether the chronological order of the status changes in the two sequences is consistent. If the chronological order of any corresponding event in the two sequences is inconsistent, the validation service triggers a dependency audit alarm, prompting the administrator to check the logical consistency between the definition of the dependency task list and the actual scheduling records.

[0058] For example, the verification service extracts the end-of-chain hash value corresponding to the task from the cloud audit database and compares it with the latest hash value in the local audit record. If the hash values ​​at both ends are completely consistent, the system further compares the state transition timestamps, determines that the verification is successful, and generates a verification success log. If the hash values ​​at both ends are inconsistent, the system determines that there is a data anomaly, triggers an alarm message and pushes it to the administrator, marking the task's audit status as abnormal and pending review. If the task has experienced a waiting dependency state and the state change sequence of the dependent tasks in the local record and the cloud record is inconsistent, the system triggers a dependency audit alarm.

[0059] This application embodiment sets the task's state to a silent review state and starts a timer through a state machine. Then, it waits for the timer to expire. When the timer expires and no interception instruction for the broadcast control task is received, the task state is transitioned from the silent review state to the pending execution state. At the same time, a local audit record is generated and stored based on a hash chain during each state transition. Meanwhile, a cloud audit chain is generated independently of the local audit record. The hash digest of the local audit record is cross-validated with the hash digest of the cloud audit chain. If the verification is inconsistent, an audit alarm is triggered. Thus, through this dual-track independent recording and cross-validation mechanism, the efficiency of the automated review process is ensured, as well as the traceability of the state transition process and the immutability of the audit data. This allows automatically approved tasks to be reliably reviewed afterward, effectively solving the security and liability determination risks existing in the simple timeout automatic approval mechanism.

[0060] Based on the same inventive concept, this application also provides a second embodiment, referring to... Figure 2 , Figure 2 This is a flowchart illustrating the second embodiment of the broadcast control task review method based on state machines in this application.

[0061] In this embodiment, the state machine-based broadcast control task review method further includes steps S21-S23: Step S21: Assess the risk of the content of the broadcast control task, and mark the risk level of the broadcast control task as Level 1, Level 2 and / or Level 3 based on the risk score; Step S22: If the risk level is Level 2, pause the timer for the silent review period and move the status of the broadcast control task to the pending confirmation state. Step S23: If the risk level is level 3, stop the timer and change the status of the broadcast control task to the blocked state.

[0062] In this embodiment, the broadcast control review system performs a risk scoring operation on the task content after the task enters a silent review state. Risk scoring involves inputting the content payload of the broadcast control task into a pre-trained natural language understanding model, which outputs a continuous risk score. The system compares this continuous score with multiple preset threshold boundaries and, based on the comparison results, labels the task as a first-level, second-level, or third-level risk, where the first level is the lowest risk level, the second level is the medium risk level, and the third level is the highest risk level. The system then performs differentiated state control operations based on the labeled risk level.

[0063] Specifically, the system inputs the text portion of the task's payload into a risk scoring model. This model, based on a transformer architecture neural network and trained on sample corpora labeled with risk levels, outputs a continuous risk score between 0 and 1 for the input text, with scores closer to 1 indicating higher content risk. After the model outputs the risk score, the system reads a preset first threshold boundary and a second threshold boundary, where the first threshold boundary is smaller than the second. The system compares the risk score with both the first and second threshold boundaries. When the risk score is less than the first threshold boundary, the system marks the task as Level 1, maintains a silent review state, and continues timing. When the risk score is greater than or equal to the first threshold boundary but less than the second threshold boundary, the system marks the task as Level 2, calls the pause interface of the timer module to stop the silent review time from accumulating, and simultaneously transitions the task state from silent review to pending confirmation via the state machine engine, pushing a risk warning message to the administrator account. In the pending confirmation state, the administrator reviews the task content and performs either a confirmation pass or a confirmation rejection. Upon confirmation, the system resumes the timer and transitions the task state back to silent review; upon rejection, the task state transitions to the blocked state. When the risk score is greater than or equal to the second threshold boundary, the system marks the task as level three. The system calls the stop interface of the timer module to destroy the timer, and at the same time, the state machine engine migrates the task status from silent review status to blocked status. The task is permanently removed from the automatic review process. Only administrators with advanced privileges can unblock the blocked task and re-enter the review process.

[0064] This embodiment introduces a risk scoring mechanism to automatically assess the security of broadcast control tasks and executes differentiated status control based on the assessment results. This achieves a balance between content security and automated review, avoiding the technical problem of high-risk content being wrongly approved due to unified automated review.

[0065] Since the system described in Embodiment 2 of this application is a system used to implement the method of Embodiment 1 of this application, those skilled in the art can understand the specific structure and variations of the system based on the method described in Embodiment 1 of this application, and therefore will not be described again here. All systems used in the method of Embodiment 1 of this application fall within the scope of protection of this application.

[0066] Based on the same inventive concept, this application also provides a third embodiment, referring to... Figure 3 , Figure 3 This is a flowchart illustrating the third embodiment of the broadcast control task review method based on state machines in this application.

[0067] In this embodiment, the state machine-based broadcast control task review method further includes steps S61-S63: Step S61: Obtain the content fingerprint of the content payload in the broadcast control task; Step S62: Perform similarity matching between the content fingerprint and each abnormal fingerprint in the abnormal content fingerprint database; Step S63: If the similarity obtained by matching is greater than or equal to the preset similarity threshold, the state of the broadcast control task is transferred to the blocked state through the state machine, and the silent review time is cleared.

[0068] In this embodiment, after acquiring a broadcast control task, before starting the silent review timer, or during the timer's operation, the broadcast control system performs intelligent security detection on the task's content payload. The system extracts a content fingerprint from the content payload. The content fingerprint is a numerical vector or hash value that represents the multimedia content's characteristics, converted using a feature extraction algorithm. The system calculates the similarity between the extracted content fingerprint and each abnormal fingerprint stored in the abnormal content fingerprint database, obtaining a set of similarity scores. The system compares the highest similarity score with a preset similarity threshold. When the highest similarity score reaches or exceeds the threshold, the system determines that the task's content is highly similar to known abnormal content, posing a security risk. It then uses a state machine to transition the task state to the blocked state and clears the silent review time. The abnormal content fingerprint database is a pre-built database storing abnormal content feature vectors. The abnormal fingerprints in the database originate from the content features of historically blocked tasks or the features of violating content actively marked by the administrator.

[0069] Specifically, the system executes different content fingerprint extraction strategies based on the data type of the content payload. When the content payload is plain text, the system uses the term frequency inverse document frequency (TNF) vectorization method, which calculates the TNF value of each word after text segmentation and combines them to form a text feature vector as the content fingerprint. When the content payload is an image, the system uses a perceptual hashing algorithm, which calculates the discrete cosine transform coefficients after scaling the image to a fixed size, extracts low-frequency coefficients to generate a fixed-length binary hash value as the content fingerprint. When the content payload is video, the system extracts the keyframe sequence of the video, calculates the perceptual hash value for each keyframe, and averages the hash values ​​of all keyframes as the video's content fingerprint. The system calculates the similarity between the extracted content fingerprint and each abnormal fingerprint in the abnormal content fingerprint database. For text feature vectors, the system calculates the cosine similarity between two vectors. For perceptual hash values, the system calculates the Hamming distance between two hash values ​​and normalizes it into a similarity score. The system selects the maximum value from all similarity scores and compares this maximum value with a preset similarity threshold. When the maximum value is greater than or equal to the preset similarity threshold, the system determines that the match is successful, immediately calls the state machine engine to transition the task state from the current state to the intercepted state, and simultaneously calls the timer module's clear interface to reset the silent review time to zero. The task is then removed from the automatic review process. When the maximum value is less than the preset similarity threshold, the system determines that the match has failed, and the task maintains its current state and continues to execute the normal process.

[0070] This application provides a state machine-based broadcast control task review device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the state machine-based broadcast control task review method in the first embodiment described above.

[0071] The following is for reference. Figure 4 This document illustrates a structural diagram of a state machine-based broadcast control task review device suitable for implementing embodiments of this application. The state machine-based broadcast control task review device in these embodiments may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Description), PMPs (Portable Media Players), and in-vehicle terminals (e.g., in-vehicle navigation terminals), as well as fixed terminals such as digital TVs and desktop computers. Figure 4The state machine-based broadcast control task review device shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0072] like Figure 4 As shown, the state machine-based broadcast control task review device may include a processing unit 1001 (e.g., a core processor, a graphics processor, etc.), which can perform various appropriate actions and processes according to programs stored in read-only memory (ROM) 1002 or programs loaded from storage device 1003 into random access memory (RAM) 1004. The random access memory 1004 also stores various programs and data required for the operation of the state machine-based broadcast control task review device. The processing unit 1001, ROM 1002, and RAM 1004 are interconnected via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to I / O interface 1006: input devices 1007 including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices 1008 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 1003 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1009. Communication device 1009 allows the state machine-based broadcast control task review equipment to communicate wirelessly or wiredly with other devices to exchange data. Although state machine-based broadcast control task review equipment with various systems is shown in the figure, it should be understood that it is not required to implement or possess all the systems shown. More or fewer systems can be implemented alternatively.

[0073] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from read-only memory 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0074] The state machine-based broadcast control task review device provided in this application, employing the state machine-based broadcast control task review method described in the above embodiments, can solve the technical problem of insufficient accuracy in reviewing sensitive content in broadcast control systems. Compared with the prior art, the beneficial effects of the state machine-based broadcast control task review device provided in this application are the same as those of the state machine-based broadcast control task review method provided in the above embodiments, and other technical features in this state machine-based broadcast control task review device are the same as those disclosed in the previous embodiment method, and will not be repeated here.

[0075] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0076] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0077] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, which are used to execute the state machine-based broadcast control task review method in the above embodiments.

[0078] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system, or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, radio frequency (RF), etc., or any suitable combination thereof.

[0079] The aforementioned computer-readable storage medium may be included in a state machine-based broadcast control task review device; or it may exist independently and not be assembled into a state machine-based broadcast control task review device.

[0080] The aforementioned computer-readable storage medium carries one or more programs. When these programs are executed by a state machine-based broadcast control task auditing device, the state machine-based broadcast control task auditing device: acquires a broadcast control task and records the silent auditing time of the broadcast control task through a state machine; if the silent auditing time is greater than or equal to a preset auditing time threshold, it transitions the state of the broadcast control task from the silent auditing state to the pending execution state through the state machine; after the state machine completes the state transition, it generates and stores local audit records of the broadcast control task based on a hash chain; it generates a cloud audit chain based on the broadcast control instructions issued from the cloud and the received execution receipts; it cross-verifies the hash digest of the local audit record with the hash digest of the cloud audit chain; and it executes the corresponding audit action based on the verification result.

[0081] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0082] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0083] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0084] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., computer programs) for executing the above-described state machine-based broadcast control task review method, thereby solving the technical problem of insufficient accuracy in reviewing sensitive content in broadcast control systems. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the state machine-based broadcast control task review method provided in the above embodiments, and will not be repeated here.

[0085] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A broadcast control task review method based on a state machine, characterized in that, The method includes the following steps: Obtain the broadcast control task and record the silent review time of the broadcast control task through a state machine; If the silent review time is greater than or equal to the preset review time threshold, the state of the broadcast control task is transitioned from the silent review state to the pending execution state through the state machine. After the state machine completes the state transition, local audit records for the broadcast control task are generated and stored based on the hash chain. Based on the broadcast control instructions issued from the cloud and the received execution receipts, a cloud audit chain is generated; The hash digest of the local audit record is cross-validated with the hash digest of the cloud audit chain, and the corresponding audit action is executed based on the verification result.

2. The broadcast control task review method based on state machine as described in claim 1, characterized in that, Before the step of acquiring the broadcast control task and setting the state of the broadcast control task to a silent review state via a state machine, the method further includes: Determine whether the content template for the broadcast control task is being used for the first time; If so, set the status of the broadcast control task to pending confirmation. Upon receiving the confirmation instruction for the broadcast control task, the content template is marked as a confirmed template, and the status of the broadcast control task is set to the silent review status.

3. The broadcast control task review method based on a state machine as described in claim 1, characterized in that, Before the step of transitioning the state of the broadcast control task from the silent review state to the pending execution state via the state machine when the silent review time is greater than or equal to a preset review time threshold, the method further includes: The content of the broadcast control task is risk-scored, and the risk level of the broadcast control task is marked as Level 1, Level 2 and / or Level 3 based on the risk score. If the risk level is Level 2, then the timer for the silent review period is paused, and the status of the broadcast control task is moved to the pending confirmation status. If the risk level is Level 3, then stop the timer and change the status of the broadcast control task to the blocked state.

4. The broadcast control task review method based on a state machine as described in claim 3, characterized in that, The step of performing a risk assessment on the content of the broadcast control task and marking the risk level of the broadcast control task as a first level, a second level, and / or a third level based on the risk assessment includes: Obtain the characteristic parameters of the current scene, including the current time period type, the historical interception rate of the user executing the broadcast control task, and the sensitivity level of the terminal involved in the broadcast control task; The risk score and the feature parameters are input into the reinforcement learning model to obtain the risk threshold boundaries of the first, second and third levels output by the reinforcement learning model. The risk level of the broadcast control task is determined based on the comparison between the risk score and the risk threshold boundary. After the broadcast control task is completed, a reward value is calculated based on the user's feedback on the broadcast control task, and the parameters of the reinforcement learning model are updated based on the reward value.

5. The broadcast control task review method based on a state machine as described in claim 1, characterized in that, Before the step of transitioning the state of the broadcast control task from the silent review state to the pending execution state via the state machine when the silent review time is greater than or equal to a preset review time threshold, the method further includes: Obtain the confirmation status identifier, risk level identifier, timer expiration identifier, and interception event identifier of the content template of the broadcast control task; The confirmation status identifier, the risk level identifier, the timer expiration identifier, and the interception event identifier are concatenated in a preset order to form a first determination vector; Using the first decision vector as input, the transition conditions of the state machine are determined; Based on the determination result, the first determination vector is combined with the current state identifier of the state machine to form a second determination vector. A hash operation is performed on the second determination vector to generate a migration certificate. The migration certificate is written into the local audit record and then the state migration is performed.

6. The broadcast control task review method based on a state machine as described in claim 1, characterized in that, The state machine-based broadcast control task review method also includes: Receive modification or interception instructions for the broadcast control task; If a modification instruction is received, the state machine is used to transition the state of the broadcast control task to the draft state, the silent review time is cleared, and the state of the broadcast control task is set to the silent review state after the broadcast control task is edited. If an interception command is received, the state machine is used to transition the state of the broadcast control task to the intercepted state, and the timer is stopped. If an emergency release instruction is received, the status of the broadcast control task will be moved to the pending execution status.

7. The broadcast control task review method based on a state machine as described in claim 1, characterized in that, After the steps of obtaining the broadcast control task and recording the silent review time of the broadcast control task through a state machine, the method further includes: Obtain the content fingerprint of the content payload in the broadcast control task; The content fingerprint is matched with each abnormal fingerprint in the abnormal content fingerprint database based on similarity. If the similarity obtained by matching is greater than or equal to the preset similarity threshold, the state of the broadcast control task is transferred to the blocked state through the state machine, and the silent review time is cleared.

8. The broadcast control task review method based on state machine as described in claim 1, characterized in that, The state machine-based broadcast control task review method also includes: Analyze the temporal dependencies between the broadcast control task and other broadcast control tasks to generate a list of dependent tasks; When there are tasks in the dependent task list of the broadcast control task that are not marked as completed or approved, the status of the broadcast control task is set to waiting for dependency, and the silent review time is cleared. Once all tasks in the dependent task list have a completed or approved status, the broadcast control task will be restored to the silent review status.

9. A broadcast control task review device based on a state machine, characterized in that, The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the state machine-based broadcast control task review method as described in any one of claims 1 to 8.

10. A storage medium, characterized in that, The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, it implements the steps of the state machine-based broadcast control task review method as described in any one of claims 1 to 8.

Images