A multi-agent collaboration-based fusion media management and control and production dual-domain linkage system and method

Abstract

The application provides a multi-agent cooperation-based fusion media management and production dual-domain linkage system and method, which comprises: a unified data center for integrating multi-element heterogeneous data and providing unified data access and calling interfaces for each agent; a central dispatch agent set for receiving and disassembling business requirements, dynamically dispatching management domain agents and production domain agents to form cross-domain task groups, and coordinating the interaction and cooperation between the agents; a management domain agent set for realizing the whole-process operation management and risk control of fusion media business; and a production domain agent set for realizing the whole-link automatic execution of fusion media content from clue mining to distribution, and maintaining real-time two-way communication with the management domain agent. The application realizes parallel cooperation of production and management, effectively reduces compliance risk, copyright risk and process lag risk, avoids large-area rework, and improves content publishing efficiency.

Description

Technical Field

[0001] This invention relates to the field of media convergence technology, and in particular to a dual-domain linkage system and method for media convergence management and production based on multi-agent collaboration. Background Technology

[0002] With the deepening of media convergence and the continuous improvement of informatization, new technologies, represented by artificial intelligence, are having an increasingly profound impact on the media industry. Currently, the industry has entered a stage of high-quality development, but traditional production and management models have accumulated many deep-seated contradictions, hindering the improvement of industry efficiency and competitiveness.

[0003] The existing data systems of media organizations are primarily production-oriented, with management-oriented data in a subordinate position. The two are independent and lack coordination. Operational data and production data have long been unable to be integrated. Operational needs can only be passed down manually at each level, resulting in lengthy processes, slow responses, and a widespread discrepancy between production arrangements and business objectives. The problem of "two separate systems for operations and production" has never been effectively resolved.

[0004] Regarding content compliance, the current "produce first, review later" model places risk control outside the production chain. There is no advance warning of compliance risks, and once a problem occurs, rework is often the only option, which is both wasteful of resources and makes it difficult to truly guarantee content security.

[0005] In terms of intelligent applications, the current introduction of AI tools is mostly localized and fragmented, with each system operating independently and lacking unified scheduling. There is also a lack of collaborative mechanisms between intelligent agents, making it impossible to form an intelligent production system covering the entire chain, and the technological potential has not been fully realized.

[0006] In addition, traditional linear production processes rely heavily on manual intervention, lack flexibility, and are difficult to adapt to the content production needs of multiple platforms and scenarios. Furthermore, data is stored in a scattered manner, and there is a lack of two-way data flow channels between production and operation.

[0007] In summary, there is an urgent need for a technical solution that can bridge the control and production domains, support multi-agent collaboration, and achieve end-to-end data connectivity, thereby fundamentally driving the transformation of the media industry's production model. Summary of the Invention

[0008] This application provides a system and method for integrated media management and production dual-domain linkage based on multi-agent collaboration, in order to solve the problems in the background art.

[0009] Other features and advantages of this application will become apparent from the following detailed description, or may be learned in part from practice of this application.

[0010] According to a first aspect of the embodiments of this application, a converged media management and production dual-domain linkage system based on multi-agent collaboration is provided, comprising: A unified data platform is used to integrate diverse and heterogeneous data and provide a unified data access and calling interface for various intelligent agents. The central scheduling intelligent agent set is used to receive and break down business requirements, dynamically schedule and control domain intelligent agents and production domain intelligent agents to form cross-domain task groups, and coordinate the interaction and cooperation between intelligent agents; A collection of intelligent agents in the control domain is used to realize the full-process operation management and risk control of converged media business; The production domain intelligent agent set is used to realize the full-link automated execution of converged media content from lead mining to distribution, and maintain real-time two-way communication with the control domain intelligent agent.

[0011] According to one embodiment of this application, the central scheduling intelligent agent set includes a demand decomposition intelligent agent, a scheduling and control intelligent agent, a cross-domain invocation intelligent agent, and an intelligent agent collaborative control intelligent agent, wherein... The demand decomposition agent is used to receive user input and perform standardized atomic decomposition according to four dimensions: business indicators, commercial constraints, delivery requirements, and time nodes, to generate atomic task packages containing task ID, task type, priority, delivery standards, and execution agent. The scheduling and control agent is used to execute a preset weighted round-robin scheduling algorithm according to the atomic task package, and to schedule agents in the production domain agent set and the control domain agent set to form a cross-domain task group; The cross-domain invocation agent is used to establish a two-way communication link between the control domain and the production domain. The aforementioned intelligent agent collaborative management agent is used to manage the interaction rules, execution sequence, and permission boundaries of each intelligent agent in each cross-domain task group.

[0012] According to one embodiment of this application, the central scheduling agent set further includes a feedback scheduling agent and a global monitoring agent, wherein... The feedback scheduling agent is used to acquire and analyze the standardized evaluation report of the published content, generate optimization suggestions, and drive the iterative update of the corresponding agent's working strategy. The global monitoring agent is used to collect the execution status, task progress and data flow information of each agent in real time, and to establish a full-link monitoring system.

[0013] According to one embodiment of this application, the control domain intelligent agent set includes an operational intelligent agent system and a control intelligent agent system. The operational intelligent agent system includes an OA intelligent agent and a financial intelligent agent, and the control intelligent agent system includes a content compliance intelligent agent and a copyright management intelligent agent. The OA intelligent agent is used to automate the cross-domain approval process; The financial intelligent agent is bound to the entire lifecycle of production tasks and is used to achieve financial control over the production process. The content compliance intelligent agent is used to achieve compliance risk management throughout the entire content production chain; The copyright management intelligent agent is used to build a full-chain copyright protection system to prevent and control copyright risks at each stage of production.

[0014] According to one embodiment of this application, the production domain intelligent agent set includes a clue collection intelligent agent, a topic planning intelligent agent, a content creation intelligent agent, a material management intelligent agent, a distribution and operation intelligent agent, and a dissemination evaluation intelligent agent, wherein... The clue-collecting intelligent agent is used for the automated collection and processing of multi-source heterogeneous content clues, generating standardized clues; The topic selection planning intelligent agent is used to generate multiple differentiated topic selection schemes and execution frameworks based on central scheduling instructions and standardized clue data, and by integrating business indicators, control rules and user profiles. The content creation intelligent agent is used to generate multi-form integrated media content in parallel based on the multiple sets of differentiated topic selection schemes and execution frameworks; The intelligent agent for material management is used for the unified management and scheduling of all categories of genuine materials, and is linked with the intelligent copyright management system. The distribution operation intelligent agent is used to construct the optimal distribution strategy based on content attributes, channel characteristics and user profiles according to the release authorization instructions of the central scheduling intelligent agent, and to release the content. The aforementioned propagation assessment agent is used for real-time collection and multi-dimensional quantitative analysis of propagation data across all channels, generating standardized assessment reports and pushing them to the feedback scheduling agent.

[0015] According to one embodiment of this application, the clue collection agent monitors hot events, public opinion dynamics and industry information across the entire network in real time, and completes the cleaning, deduplication, classification and quantitative evaluation of clues to generate standardized clues.

[0016] According to one embodiment of this application, the content creation intelligent agent simultaneously completes text and image writing, audio and video editing, special effects packaging, and multi-terminal version optimization based on the topic selection plan, and receives the verification results of the control domain in real time and dynamically adjusts the content.

[0017] According to one embodiment of this application, the material management intelligent agent is equipped with a semantic tagging system to achieve intelligent retrieval, and it works in conjunction with the content creation intelligent agent in the content creation stage to automatically match the optimal materials.

[0018] According to one embodiment of this application, the unified data platform adopts a lake-warehouse integrated architecture, integrating the Hadoop distributed file system and the Flink stream computing engine; the unified data platform also deploys a full-process data traceability probe, which is used to collect operational data, status data and result data of each stage of production and control.

[0019] According to a second aspect of the embodiments of this application, a method for media convergence control and production dual-domain linkage based on multi-agent collaboration is provided, implemented based on the media convergence control and production dual-domain linkage system based on multi-agent collaboration described in the first aspect, specifically including: System initialization: Build a unified data platform, complete the connection and adaptation of various business systems in the control domain and production domain, deploy the central scheduling intelligent agent set, the control domain intelligent agent set and the production domain intelligent agent set, and configure basic linkage rules, permission system and process templates; Demand Triggering and Decomposition: The demand decomposition intelligent agent receives user input demands and performs standardized atomic decomposition, generates atomic task packages, and pushes operational demands to the operational intelligent agent system and management demands to the management intelligent agent system. Cross-domain task group formation: The scheduling and control agent dynamically forms a cross-domain agent task group based on the decomposed atomic task package. The cross-domain calling agent opens the dual-domain communication link and injects operational constraints and control rules. The agent and the control agent clarify the interaction sequence and permissions. Parallel production and real-time control: The intelligent agent in the production domain executes the entire content production process according to the workflow, while the intelligent agent in the control domain synchronously completes compliance verification, copyright verification, OA approval and financial control through cross-domain call mechanism; Cross-domain parallel final review and distribution: The final draft of the content in the production domain and the verification results in the control domain are synchronized to the global monitoring intelligent agent, and the central scheduling intelligent agent initiates the cross-domain parallel final review; if the review is passed, a distribution instruction is triggered, and the distribution operation intelligent agent completes the content release across all channels; if the review fails, the feedback scheduling intelligent agent initiates the content correction process. Propagation assessment and closed-loop optimization: The propagation assessment agent collects and analyzes propagation data from all channels and generates a standardized assessment report; the feedback scheduling agent combines the standardized assessment report with full-process traceability data to generate optimization suggestions and pushes them back to each agent, driving the system's self-evolution.

[0020] Compared with existing technologies, the beneficial effects of adopting the above technical solution are as follows: 1. A unified data platform was built, which integrated multi-source heterogeneous data, eliminated data silos, and provided a unified, standardized, and efficient data support foundation for the dual-domain linkage and multi-agent collaboration between the control and production domains, realizing the interconnection and sharing of data across the entire domain.

[0021] 2. A three-layer intelligent agent architecture of "central scheduling - control domain - production domain" was established, clarifying the functional boundaries and interaction rules of each intelligent agent. Through central scheduling, dynamic networking, collaborative operation and load balancing of intelligent agents were realized, which significantly improved the overall coordination and operating efficiency of the system.

[0022] 3. It has achieved parallel collaboration between production and control, embedding compliance review, copyright verification, OA approval and financial control into the entire content production process, transforming post-event control into full-process parallel control, effectively reducing compliance risks, copyright risks and process delay risks, avoiding large-scale rework and improving content release efficiency.

[0023] 4. A full-process data traceability and closed-loop optimization mechanism was established. Through data traceability probes, the entire business chain was made traceable and auditable. Combined with a multi-touchpoint attribution model based on Shapley values ​​and a feedback scheduling mechanism, the continuous iteration of the agent's working strategy and the self-evolution of system capabilities were realized.

[0024] 5. It has achieved automated execution of the entire converged media business chain, greatly reducing manual intervention, lowering human error and operating costs, enabling rapid response to content production needs across multiple platforms, scenarios, and formats, and significantly enhancing the overall competitiveness of the converged media platform. Attached Figure Description

[0025] 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. It is obvious that the drawings described below are merely some embodiments of this application, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0026] Figure 1 This is a schematic diagram of a multi-agent collaborative media management and production dual-domain linkage system according to an embodiment of this application.

[0027] Figure 2 This is a flowchart of a method for multi-agent collaborative media management and production dual-domain linkage based on an embodiment of this application.

[0028] Figure 3 This is a schematic diagram of the structure of an electronic device according to an embodiment of this application. Detailed Implementation

[0029] The embodiments of this application are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar modules or modules having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain this application, and should not be construed as limiting this application. Rather, the embodiments of this application include all variations, modifications, and equivalents falling within the spirit and scope of the appended claims.

[0030] To address the shortcomings of existing technologies, this application provides a multi-agent collaborative media management and production dual-domain linkage system. By constructing a unified data platform and a three-layer agent architecture, it achieves deep integration and real-time linkage between the management and production domains, establishing an intelligent operation system covering the entire lifecycle of "operation-management-production-dissemination-feedback", fundamentally improving the production efficiency, management capabilities, and dissemination effects of the media convergence platform.

[0031] Please refer to Figure 1 This integrated media management and production dual-domain linkage system based on multi-agent collaboration mainly includes a unified data platform, a central scheduling agent set, a management domain agent set, and a production domain agent set.

[0032] Specifically, the unified data platform is mainly used to integrate diverse and heterogeneous data and provide a unified data access and calling interface for each intelligent agent. In this embodiment, the unified data platform realizes the physical and logical foundation for dual-domain linkage. By constructing a fusion data view covering all dimensions of operation, management, production, and dissemination, it provides a unified data access and calling interface for each intelligent agent and supports the collection, storage, processing, and traceability of data throughout the entire process.

[0033] In one embodiment, the unified data platform adopts a lakeware architecture, integrating the Hadoop Distributed File System and the Flink stream processing engine to achieve batch processing of offline data and low-latency processing of real-time data, providing a unified GraphQL query interface. Furthermore, the unified data platform also deploys end-to-end data traceability probes to collect operational, status, and result data from each stage of production and control, enabling full-chain traceability and auditability of the business.

[0034] By integrating multi-source heterogeneous data through a unified data platform, data silos are eliminated, providing a unified, standardized, and efficient data support foundation for dual-domain linkage and multi-agent collaboration between the control and production domains, and realizing the interconnection and sharing of data across the entire domain.

[0035] In this embodiment, the central scheduling agent set serves as the global collaborative control center of the system. It is mainly used to receive and break down business requirements, dynamically schedule and manage domain agents and production domain agents to form cross-domain task groups, and coordinate the interaction and cooperation between agents.

[0036] Specifically, in this embodiment, the central scheduling intelligent agent set includes a demand decomposition intelligent agent, a scheduling and control intelligent agent, a cross-domain invocation intelligent agent, and an intelligent agent collaboration and control intelligent agent. It should be noted that the intelligent agents included in the central scheduling intelligent agent set here are only illustrative examples and can be added or removed according to actual functional requirements.

[0037] For the demand decomposition agent, it has a large language model that is finely tuned for the media business domain. It is mainly used to receive input information such as user needs, publicity instructions, and hot topic clues. It performs standardized atomic decomposition according to four dimensions: business indicators, business constraints, delivery requirements, and time nodes, and generates atomic task packages containing task ID, task type, priority, delivery standards, and execution agent.

[0038] For scheduling and control agents, it is mainly used to execute a preset weighted round-robin scheduling algorithm, which comprehensively considers the real-time load of each agent, the success rate of historical tasks, and the matching degree of task types to perform task allocation and load balancing, so as to realize the on-demand networking and collaborative operation of agents.

[0039] Cross-domain intelligent agents are mainly used to establish a two-way communication link between the control domain and the production domain, enabling cross-domain calls and data interaction between control and production intelligent agents, and ensuring that production and control processes proceed in parallel.

[0040] For intelligent agents in collaborative management, the main approach is based on the SAGA transaction coordination model. Each cross-domain task group is assigned a globally unique TraceID, which clarifies the interaction rules, execution sequence, and permission boundaries of each intelligent agent. This approach coordinates and resolves conflicts during the collaborative process of intelligent agents, ensuring the orderly operation of the entire process.

[0041] Furthermore, the central scheduling agent set is also used to monitor task execution status and drive full-process optimization based on feedback data. Based on this, the central scheduling agent set further includes feedback scheduling agents and global monitoring agents. The feedback scheduling agent is mainly used to acquire propagation effect evaluation data and full-process traceability data, perform structured analysis to generate optimization suggestions, and push them back to the corresponding agents to drive iterative updates of the agents' work strategies. The global monitoring agent is mainly used to collect the execution status, task progress, and data flow information of each agent in real time, establish a full-link monitoring system, provide real-time alarms and emergency handling for abnormal situations, and ensure stable system operation and full business traceability.

[0042] In this embodiment, the set of intelligent agents in the control domain is mainly used to realize the full-process operation management and risk control of converged media business. Specifically, the set of intelligent agents in the control domain is divided into an operational intelligent agent system and a control intelligent agent system. The operational intelligent agent system includes an OA intelligent agent and a financial intelligent agent, while the control intelligent agent system includes a content compliance intelligent agent and a copyright management intelligent agent.

[0043] For OA intelligent agents, they are mainly used to realize the automated flow of cross-domain approval processes, receive instructions from the central scheduling intelligent agent, automatically trigger corresponding approval nodes in sync with the production task progress, and realize the real-time transparent transmission and closed-loop execution of approval instructions from the control domain to the production domain.

[0044] For the financial intelligent agent, this agent is bound to the entire lifecycle of production tasks. It is mainly used to realize the pre-control of project budget, automatic accounting of production process costs, automatic settlement of expenses and real-time statistics of revenue data, and automatically complete budget deduction and cost collection according to the production progress.

[0045] The content compliance intelligent agent is mainly used to achieve compliance risk control throughout the entire content production chain. It supports real-time streaming verification of fine-grained content such as text paragraphs and audio / video scenes. The verification results are sent back to the corresponding production stage in milliseconds, enabling proactive identification and immediate correction of compliance risks.

[0046] The copyright management intelligent agent is mainly used to build a full-chain copyright protection system. It automatically completes the pre-verification of authorization legality in the material retrieval stage, performs originality detection in the content generation stage, and embeds an tamper-proof copyright traceability mark in the content publishing stage, so as to achieve full-process prevention and control of copyright risks.

[0047] In this embodiment, the production domain intelligent agent set is mainly used to automate the entire process of converged media content execution from lead generation to distribution, and maintains real-time two-way communication with the control domain intelligent agent to ensure that the entire production process complies with operational requirements and control rules. Specifically, the production domain intelligent agent set includes a lead gathering intelligent agent, a topic planning intelligent agent, a content creation intelligent agent, a material management intelligent agent, a distribution and operation intelligent agent, and a dissemination evaluation intelligent agent. Similar to the central scheduling intelligent agent set, the intelligent agents included in the production domain intelligent agent set here are only for illustrative purposes and can be added or removed according to actual functional requirements.

[0048] The intelligent agent for clue collection is mainly used for the automated collection and processing of multi-source heterogeneous content clues, real-time monitoring of hot events, public opinion dynamics and industry information across the entire network, completion of clue cleaning, deduplication, classification and quantitative evaluation of heat, and push of standardized clues to subsequent production stages.

[0049] The topic selection and planning intelligent agent is mainly used to receive central dispatch instructions and clue data, integrate business indicators, control rules and user profiles, and use multi-objective optimization algorithms to generate multiple sets of differentiated topic selection schemes and execution frameworks, supporting manual review and adjustment.

[0050] For the content creation intelligent agent, it is mainly used for the parallel generation and adaptation of multi-form converged media content. According to the topic selection plan, it can simultaneously complete the writing of graphics and text, audio and video editing, special effects packaging and multi-terminal version optimization, receive the verification results of the control domain in real time and dynamically adjust the content, so as to realize the creation, verification and optimization at the same time.

[0051] The intelligent agent for material management is mainly used for the unified management and scheduling of all categories of genuine materials. It builds a semantic tagging system to achieve intelligent retrieval, automatically matches the best materials in the content creation process, and works in conjunction with the intelligent agent for copyright management to complete authorization pre-verification.

[0052] For the distribution operation intelligent agent, it is mainly used to receive the release authorization instructions from the central dispatch, build the optimal distribution strategy based on content attributes, channel characteristics and user profiles, and realize one-click distribution across all channels, scheduled release and real-time monitoring and feedback of release status.

[0053] For the intelligent agent for communication assessment, it is mainly used for real-time collection and multi-dimensional quantitative analysis of communication data across all channels. Based on the Shapley value multi-touchpoint attribution model, it calculates the marginal communication contribution of each channel, generates a standardized assessment report, and pushes it to the feedback scheduling stage to drive the optimization of the entire process.

[0054] The multi-agent collaborative media management and production dual-domain linkage system proposed in this embodiment establishes a three-layer agent architecture of "central scheduling - management domain - production domain", clarifies the functional boundaries and interaction rules of each agent, and realizes dynamic networking, collaborative operation and load balancing of agents through the central scheduling agent, which significantly improves the overall coordination and operating efficiency of the system.

[0055] To further illustrate the working principle of the multi-agent collaborative media control and production dual-domain linkage system in this embodiment, this embodiment also proposes a media control and production dual-domain linkage method based on the aforementioned multi-agent collaborative media control and production dual-domain linkage system. Please refer to [link / reference]. Figure 2 Specifically, it includes the following steps: S100 System Initialization: Build a unified data platform, complete the connection and adaptation of various business systems in the control domain and production domain, deploy the central scheduling intelligent agent set, the control domain intelligent agent set, and the production domain intelligent agent set, and configure basic linkage rules, permission system, and process templates.

[0056] In this embodiment, the unified data platform adopts a Lakehouse architecture, integrating the Hadoop Distributed File System and the Flink stream computing engine to achieve deep integration of heterogeneous business systems through an automated integration process. In practical applications, after logging into the system, administrators input the communication protocol parameters, API keys, and database access permissions of the systems to be integrated through the configuration interface. The system then automatically executes the following construction logic: Protocol handshake and link establishment: Utilize RESTful API, gRPC, or Enterprise Service Bus (ESB) to automatically establish a two-way communication tunnel with the production domain (media asset system, broadcasting system, China Blue Cloud converged media technology platform, AIGC and digital human production platform, etc.) and the control domain (finance, advertising, OA, copyright system, etc.).

[0057] Heterogeneous data standardization: The system automatically cleans and transforms the collected heterogeneous data, mapping it uniformly into a standard metadata format that can be recognized by the digital intelligence management service system.

[0058] Access Control and Probe Deployment: A comprehensive service management system is built, covering all critical user operation permissions throughout the service lifecycle. The system automatically deploys full-process data traceability probes at each interface node to ensure that every production and control step has corresponding access control and status monitoring capabilities.

[0059] The platform management panel provides a centralized monitoring view, enabling administrators to view the platform's operational status, computing load, and data flow in real time. Through pre-configured structured workflow templates for routine production, customized commercial applications, and emergency reporting, seamless integration from the underlying data environment to the upper-level business logic is achieved.

[0060] The aforementioned construction logic can be executed through existing scripts or other methods. However, the execution method is not the focus of this application, and existing solutions can be used. Therefore, it will not be elaborated here.

[0061] S200, Demand Triggering and Decomposition: The demand decomposition intelligent agent receives user input demands and performs standardized atomic decomposition to generate atomic task packages. It then pushes operational demands to the operational intelligent agent system and control demands to the control intelligent agent system.

[0062] After logging in, users input their specific production needs through a conversational interface. The system then uses a large language model finely tuned from corpora in the broadcasting industry to transform vague user intentions into standardized atomic task packages with operational constraints, control rules, and production goals through natural language processing (NLP) technology.

[0063] For example, if a user inputs a production requirement of "creating a short video about flood relief efforts in a certain area, requiring it to include the platform logo, be released exclusively online, and have a controlled budget," the intelligent agent, upon receiving the instruction, initiates a multi-dimensional semantic parsing process: First, it uses entity recognition technology to pinpoint the core topic of "flood relief"; second, it automatically associates the video with pre-defined compliance guidelines for disaster reporting within the system. The system then breaks down the overall task into interconnected atomized subtasks and defines attribute tags for each subtask: production attributes (such as storyboard editing, digital human compositing parameters), operational attributes (such as the traffic acquisition budget for the first release channel, project cost limits), and control attributes (such as copyright ownership verification, security review level). The decomposed task chain generates a visual logical topology diagram in real time in the background, which the administrator can preview and click "confirm execution," thus completing the formal transformation from user idea to system instruction.

[0064] S300, Cross-Domain Task Group Formation: The scheduling and control agent dynamically forms a cross-domain agent task group based on the decomposed atomic task package. The cross-domain calling agent opens the dual-domain communication link and injects operational constraints and control rules. The agent collaboratively manages the agent to clarify the interaction sequence and permissions.

[0065] In this embodiment, the scheduling and control agent constructs a directed acyclic graph (DAG) based on the causal dependencies between tasks and uses a distributed heartbeat monitoring mechanism to achieve dynamic resource allocation based on the real-time load of nodes. Specifically, after the scheduling and control agent detects the "confirm execution" task signal, it first determines the topological order of task execution (e.g., copyright pre-check must be passed before material download can start). The system automatically scans the real-time CPU / memory load and I / O throughput of each execution node in the current cloud environment, and uses an improved weighted round-robin (WRR) algorithm to select the most suitable instance from the agent resource pool (e.g., calling the high-performance rendering agent to process video). During the cross-domain link establishment phase, the cross-domain calling agent injects the compliance keyword list, budget cap, and other control rules into the temporary running memory of the production agent through the event bus. The agent and the control agent synchronously allocate globally unique TraceIDs. Administrators can observe the "calling" status and task mounting status of each agent node in real time through a centralized dashboard to ensure that the rule benchmarks of each link are aligned before startup.

[0066] S400, Parallel Production and Real-time Control: The production domain's intelligent agents execute the entire content production process according to the workflow, while the control domain's intelligent agents synchronously complete compliance verification, copyright verification, OA approval, and financial control through cross-domain invocation mechanisms.

[0067] In this embodiment, after the production domain agent initiates the content creation process, the control domain agent simultaneously starts "accompanying" streaming monitoring, realizing parallel linkage between production and control. This parallel production and real-time control process abandons the traditional "offline submission" mode and utilizes zero-copy technology and asynchronous data tunnels to achieve "synchronous and heterogeneous" interaction between production data streams and control verification streams.

[0068] Specifically, when a production domain agent (such as an AIGC video rendering or digital human synthesis agent) produces intermediate materials (such as storyboards or rendering frames) during the generation process, the system pushes the data copy to the verification queue of the control domain in real time through a kernel-level data mirroring mechanism.

[0069] In this embodiment, specific control measures are given: (1) Parallel multidimensional verification: The compliance agent uses a multimodal deep learning algorithm to perform visual identification comparison on each frame of the picture, and performs ASR semantic filtering on the audio stream. (2) Instant feedback of conflicts: If it is found that the content deviates from the guidance or uses unauthorized materials, the compliance agent immediately sends a "logic interception interruption" signal through the central feedback loop. (3) SAGA transaction compensation: After receiving the signal, the collaborative control agent executes the SAGA distributed transaction compensation logic according to TraceID: The system automatically records the current computing power cost that has been lost, directs the production node to perform "state rollback", and automatically drives the material agent to match the compliance alternative solution, guiding the creation node to regenerate the content.

[0070] The entire process runs in an automated closed loop in the background. Administrators only need to view the trajectory logs that are automatically corrected by the system, achieving millisecond-level dynamic alignment between the production process and compliance control, which greatly reduces the risk of redoing after large-scale rendering.

[0071] S500, Cross-Domain Parallel Final Review and Distribution: The final draft of the content in the production domain and the verification results in the control domain are synchronized to the global monitoring intelligent agent, and the central scheduling intelligent agent initiates the cross-domain parallel final review; if the review is passed, the distribution instruction is triggered, and the distribution operation intelligent agent completes the content release across all channels; if the review is not passed, the feedback scheduling intelligent agent initiates the content correction process.

[0072] In this embodiment, for the final review and distribution link, an immutable business delivery environment is constructed by using asymmetric encryption technology and a Merkle tree hash chain verification scheme.

[0073] After the produced content is packaged, the central scheduling agent retrieves the end-to-end TraceID records and triggers a parallel final review. In this embodiment, the final review content includes: (1) The system automatically collects the digital fingerprint (SHA-256 hash value) of the finished product and compares it with the middleware hash record that has passed the verification in S400 to confirm the integrity of the content.

[0074] (2) Multidimensional state aggregation: The financial agent retrieves the ERP interface to verify the budget reconciliation results, the copyright agent generates a unique copyright certificate, and the compliance agent issues a final audit report with an electronic signature.

[0075] (3) Token issuance and encryption: The system generates a time-sensitive issuance authorization token only when the aforementioned verification results are all within the preset threshold and logically matched.

[0076] The distribution and operation intelligent agent can only activate the downlink CDN acceleration interface and execute one-click concurrent delivery across all touchpoints after verifying the validity of the token signature. If the system detects an anomaly (such as a 3% budget overrun), it will accurately locate the faulty node and pop up a "Manual Intervention Review" dialog box in the administrator backend, allowing administrators to remotely confirm or correct the content with one click via mobile device, ensuring that every second of broadcast content has full control and authorization.

[0077] S600, Propagation Assessment and Closed-Loop Optimization: The propagation assessment agent collects and analyzes propagation data from all channels and generates a standardized assessment report; the feedback scheduling agent combines the standardized assessment report with full-process traceability data to generate optimization suggestions and pushes them back to each agent, driving the system's self-evolution.

[0078] In this embodiment, distributed crawlers and open APIs of various platforms are used to collect traffic data in real time, channel value is attributed based on the Shapley value model, and the evaluation report is fed back to the base as a system feedback instruction.

[0079] Specifically, after content goes live, the dissemination evaluation agent automatically monitors the dissemination effectiveness across all terminals (Douyin, WeChat, Weibo, and other multiple platforms), including completion rate, user interaction rate, and conversion rate. The system utilizes intelligent analysis tools to deeply mine the collected data, calculating the marginal contribution of each distribution channel to the dissemination target. The analysis report is not only used for business decisions but also serves as a closed-loop optimization parameter, influencing the demand decomposition agent and the scheduling and control agent. Through this automated workflow design, the system can automatically adjust the agent selection weights and budget allocation models for similar future tasks based on historical dissemination performance. Administrators can intuitively observe the step-by-step improvement in system operational efficiency through evolution logs, thereby achieving self-iteration and continuous capability enhancement of the converged media production control system.

[0080] Through the complete execution process from S100 to S600, the system achieves a comprehensive workflow, from triggering user natural language requests, atomically breaking down instructions, and load-aware cross-domain dynamic networking, to "accompanying" real-time compliance verification based on zero-copy technology, then to cross-domain parallel final review and distribution based on hash fingerprints and encrypted tokens, and finally to propagation effect evaluation and closed-loop feedback of the entire chain logic based on the Shapley value model. Administrators can intuitively observe the step-by-step improvement in system operational efficiency through the evolution log, thereby realizing the self-iteration and continuous enhancement of the converged media production control system.

[0081] Based on the same technical concept, embodiments of this application also provide an electronic device that can implement the multi-agent collaborative media management and production dual-domain linkage system or method provided in the above embodiments of the present invention. In one embodiment, the electronic device can be a server, a terminal device, or other electronic devices. Figure 3 As shown, the electronic device may include: At least one processor and a memory connected to the at least one processor. In this embodiment of the invention, the specific connection medium between the processor and the memory is not limited. Figure 3 The example used is the connection between the processor and memory via a bus. The bus... Figure 3 The connections between other components are indicated by thick lines and are for illustrative purposes only, not as limiting information. Buses can be divided into address buses, data buses, control buses, etc., but for ease of representation, [the specific bus type is not shown here]. Figure 3 The processor is represented by a single thick line, but this does not imply that there is only one bus or one type of bus. Alternatively, a processor can also be called a controller; there are no restrictions on the name.

[0082] In this embodiment of the invention, the memory stores instructions executable by at least one processor. By executing the instructions stored in the memory, the at least one processor can execute the aforementioned multi-agent collaborative media control and production dual-domain linkage system or method. The processor can implement... Figure 3 The functions of each module in the device shown.

[0083] The processor is the control center of the device. It can connect to various parts of the control device through various interfaces and lines. By running or executing instructions stored in memory and calling data stored in memory, it can monitor the device's various functions and process data, thereby enabling overall monitoring of the device.

[0084] In an alternative design, the processor may include one or more processing units. The processor may integrate an application processor and a modem processor, wherein the application processor primarily handles the operating system, user interface, and applications, while the modem processor primarily handles wireless communication. It is understood that the modem processor may also not be integrated into the processor. In some embodiments, the processor and memory may be implemented on the same chip; in some embodiments, they may also be implemented separately on separate chips.

[0085] The processor can be a general-purpose processor, such as a CPU, digital signal processor, application-specific integrated circuit, field-programmable array, or other programmable logic device, discrete gate or transistor logic device, or discrete hardware component, capable of implementing or executing the methods, steps, and logic block diagrams disclosed in the embodiments of this invention. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the multi-agent collaborative media control and production dual-domain linkage system or method disclosed in the embodiments of this invention can be directly manifested as execution by a hardware processor, or execution by a combination of hardware and software modules within the processor.

[0086] Memory, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs, non-volatile computer-executable programs, and modules. Memory can include at least one type of storage medium, such as flash memory, hard disk, multimedia card, card-type memory, random access memory (RAM), static random access memory (SRAM), programmable read-only memory (PROM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), magnetic memory, magnetic disk, optical disk, etc. Memory is any other medium capable of carrying or storing desired program code in the form of instructions or data structures, and accessible by a computer, but is not limited thereto. In embodiments of the present invention, memory can also be a circuit or any other device capable of implementing storage functions, used to store program instructions and / or data.

[0087] By designing and programming the processor, the code corresponding to the multi-agent collaborative media management and production dual-domain linkage system or method described in the foregoing embodiments can be embedded into the chip, enabling the chip to execute the steps of the methods described in the foregoing embodiments during runtime. How to design and program the processor is a technique well-known to those skilled in the art and will not be elaborated upon here.

[0088] Based on the same inventive concept, embodiments of the present invention also provide a storage medium storing computer instructions, which, when executed on a computer, cause the computer to execute the aforementioned multi-agent collaborative media control and production dual-domain linkage system or method.

[0089] In some alternative embodiments, the present invention also provides a multi-agent collaborative media convergence control and production dual-domain linkage system or method that can also be implemented as a program product, which includes program code. When the program product is run on a device, the program code is used to cause the control device to perform the steps in the multi-agent collaborative media convergence control and production dual-domain linkage system or method according to various exemplary embodiments of the present invention as described in this specification.

[0090] It should be noted that although several units or sub-units of the apparatus have been mentioned in the detailed description above, this division is merely exemplary and not mandatory. In fact, according to embodiments of the invention, the features and functions of two or more units described above can be embodied in one unit. Conversely, the features and functions of one unit described above can be further divided and embodied by multiple units. Furthermore, although the operation of the method of the invention is described in a specific order in the drawings, this does not require or imply that these operations must be performed in that specific order, or that all the operations shown must be performed to achieve the desired result. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined into one step, and / or one step may be broken down into multiple steps.

[0091] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied 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.

[0092] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. 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 server, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0093] Program code for performing the operations of this invention can be written using any combination of one or more programming languages, including object-oriented programming languages ​​such as Java and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computing device, partially on the user's device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0094] In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0095] 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.

[0096] 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.

[0097] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A multi-agent collaborative media management and production dual-domain linkage system, characterized in that, include: A unified data platform is used to integrate diverse and heterogeneous data and provide a unified data access and calling interface for various intelligent agents. The central scheduling intelligent agent set is used to receive and break down business requirements, dynamically schedule and control domain intelligent agents and production domain intelligent agents to form cross-domain task groups, and coordinate the interaction and cooperation between intelligent agents; A collection of intelligent agents in the control domain is used to realize the full-process operation management and risk control of converged media business; The production domain intelligent agent set is used to realize the full-link automated execution of converged media content from lead mining to distribution, and maintain real-time two-way communication with the control domain intelligent agent.

2. The multi-agent collaborative media management and production dual-domain linkage system according to claim 1, characterized in that, The central scheduling intelligent agent set includes demand decomposition intelligent agents, scheduling and control intelligent agents, cross-domain invocation intelligent agents, and intelligent agent collaborative control intelligent agents, among which... The demand decomposition agent is used to receive user input and perform standardized atomic decomposition according to four dimensions: business indicators, commercial constraints, delivery requirements, and time nodes, to generate atomic task packages containing task ID, task type, priority, delivery standards, and execution agent. The scheduling and control agent is used to execute a preset weighted round-robin scheduling algorithm according to the atomic task package, and to schedule agents in the production domain agent set and the control domain agent set to form a cross-domain task group; The cross-domain invocation agent is used to establish a two-way communication link between the control domain and the production domain. The aforementioned intelligent agent collaborative management agent is used to manage the interaction rules, execution sequence, and permission boundaries of each intelligent agent in each cross-domain task group.

3. The multi-agent collaborative media management and production dual-domain linkage system according to claim 1 or 2, characterized in that, The central scheduling agent set also includes a feedback scheduling agent and a global monitoring agent, wherein... The feedback scheduling agent is used to acquire and analyze the standardized evaluation report of the published content, generate optimization suggestions, and drive the iterative update of the corresponding agent's working strategy. The global monitoring agent is used to collect the execution status, task progress and data flow information of each agent in real time, and to establish a full-link monitoring system.

4. The multi-agent collaborative media management and production dual-domain linkage system according to claim 1, characterized in that, The set of intelligent agents in the control domain includes an operational intelligent agent system and a control intelligent agent system. The operational intelligent agent system includes an OA intelligent agent and a financial intelligent agent, and the control intelligent agent system includes a content compliance intelligent agent and a copyright management intelligent agent. The OA intelligent agent is used to automate the cross-domain approval process; The financial intelligent agent is bound to the entire lifecycle of production tasks and is used to achieve financial control over the production process. The content compliance intelligent agent is used to achieve compliance risk management throughout the entire content production chain; The copyright management intelligent agent is used to build a full-chain copyright protection system to prevent and control copyright risks at each stage of production.

5. The multi-agent collaborative media management and production dual-domain linkage system according to claim 3, characterized in that, The set of production domain intelligent agents includes intelligence agents for clue gathering, topic planning, content creation, material management, distribution and operation, and dissemination evaluation. The clue-collecting intelligent agent is used for the automated collection and processing of multi-source heterogeneous content clues, generating standardized clues; The topic selection planning intelligent agent is used to generate multiple differentiated topic selection schemes and execution frameworks based on central scheduling instructions and standardized clue data, and by integrating business indicators, control rules and user profiles. The content creation intelligent agent is used to generate multi-form integrated media content in parallel based on the multiple sets of differentiated topic selection schemes and execution frameworks; The intelligent agent for material management is used for the unified management and scheduling of all categories of genuine materials, and is linked with the intelligent copyright management system. The distribution operation intelligent agent is used to construct the optimal distribution strategy based on content attributes, channel characteristics and user profiles according to the release authorization instructions of the central scheduling intelligent agent, and to release the content. The aforementioned propagation assessment agent is used for real-time collection and multi-dimensional quantitative analysis of propagation data across all channels, generating standardized assessment reports and pushing them to the feedback scheduling agent.

6. The multi-agent collaborative media management and production dual-domain linkage system according to claim 5, characterized in that, The intelligent agent for collecting clues monitors trending events, public opinion dynamics, and industry news across the entire network in real time, and completes the cleaning, deduplication, classification, and quantitative evaluation of the heat of clues to generate standardized clues.

7. The multi-agent collaborative media management and production dual-domain linkage system according to claim 5, characterized in that, The content creation agent simultaneously completes text and image writing, audio and video editing, special effects packaging, and multi-terminal version optimization according to the topic selection plan, and receives the verification results of the control domain in real time and dynamically adjusts the content.

8. The multi-agent collaborative media management and production dual-domain linkage system according to claim 5, characterized in that, The material management intelligent agent is equipped with a semantic tagging system to achieve intelligent retrieval. It works in conjunction with the content creation intelligent agent during the content creation process to automatically match the optimal materials.

9. The multi-agent collaborative media management and production dual-domain linkage system according to claim 1, characterized in that, The unified data platform adopts a lake-warehouse integrated architecture, integrating the Hadoop distributed file system and the Flink stream computing engine; the unified data platform also deploys a full-process data traceability probe, which is used to collect operational data, status data and result data of each stage of production and control.

10. A method for integrated media management and production domain linkage based on multi-agent collaboration, characterized in that, The system is implemented based on any one of claims 1 to 9, specifically including: System initialization: Build a unified data platform, complete the connection and adaptation of various business systems in the control domain and production domain, deploy the central scheduling intelligent agent set, the control domain intelligent agent set and the production domain intelligent agent set, and configure basic linkage rules, permission system and process templates; Demand Triggering and Decomposition: The demand decomposition intelligent agent receives user input demands and performs standardized atomic decomposition, generates atomic task packages, and pushes operational demands to the operational intelligent agent system and management demands to the management intelligent agent system. Cross-domain task group formation: The scheduling and control agent dynamically forms a cross-domain agent task group based on the decomposed atomic task package. The cross-domain calling agent opens the dual-domain communication link and injects operational constraints and control rules. The agent and the control agent clarify the interaction sequence and permissions. Parallel production and real-time control: The intelligent agent in the production domain executes the entire content production process according to the workflow, while the intelligent agent in the control domain synchronously completes compliance verification, copyright verification, OA approval and financial control through cross-domain call mechanism; Cross-domain parallel final review and distribution: The final draft of the content in the production domain and the verification results in the control domain are synchronized to the global monitoring intelligent agent, and the central scheduling intelligent agent initiates the cross-domain parallel final review; if the review is passed, a distribution instruction is triggered, and the distribution operation intelligent agent completes the content release across all channels; if the review fails, the feedback scheduling intelligent agent initiates the content correction process. Propagation assessment and closed-loop optimization: The propagation assessment agent collects and analyzes propagation data from all channels and generates a standardized assessment report; the feedback scheduling agent combines the standardized assessment report with full-process traceability data to generate optimization suggestions and pushes them back to each agent, driving the system's self-evolution.

Images