Intelligent workflow interaction configuration method and system based on large model

Abstract

The invention discloses an intelligent workflow interaction configuration method and system based on a large model, and belongs to the technical field of workflow configuration, and the method specifically comprises the steps: obtaining function realization demand information of a user, and combining a plurality of preset models, a knowledge base and plug-ins to construct an overall large model; the large model can output corresponding output demand information based on the input demand information, and supports knowledge retrieval and configuration; in the construction process, the steps of model selection, knowledge base screening, plug-in integration, parameter adjustment and the like are involved; and the output process comprises generation, evaluation, decision analysis and optimization selection of a workflow configuration scheme. According to the invention, intelligence and automation of workflow configuration are realized, and the configuration efficiency and accuracy are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of workflow configuration, and in particular to an intelligent workflow interaction configuration method and system based on a large model. Background Art

[0002] With the continuous development of corporate business and the advancement of digital transformation, workflow management systems play a vital role in corporate operations. Traditional workflow configuration methods often require professional technicians to define and set up through complex programming or the use of specific workflow modeling tools. This not only requires configuration personnel to have deep technical knowledge, but also the entire configuration process is cumbersome and time-consuming, making it difficult to quickly respond to changes in business needs. In addition, it is almost impossible for business personnel with non-technical backgrounds to understand and participate in the configuration of workflows, resulting in a communication gap between business and technology, which further affects the optimization and adjustment efficiency of workflows.

[0003] For example, a Chinese patent with authorization announcement number CN112148416B discloses an information interaction method, device, equipment and storage medium, including: when a mobile interactive device establishes a communication connection with a target scene device, determining the interactive authentication information corresponding to the target scene device; based on the interactive authentication information, obtaining the interactive workflow information corresponding to the target scene device, and displaying the interactive workflow information on the display interface of the mobile interactive device. The technical solution of this embodiment can flexibly configure the interactive relationship between the mobile interactive device and the target scene device, and is particularly suitable for interacting with multiple target scene devices through one mobile interactive device. Moreover, based on the interactive authentication information, the interactive workflow information corresponding to the target scene device is obtained, and the interactive workflow information is displayed, which can facilitate users to view and operate the interactive workflow information, and realize the interaction between the mobile interactive device and the target scene device in a targeted manner.

[0004] For example, a Chinese patent with the authorization announcement number CN117406979B discloses a method and system for designing an interface interaction for a computational workflow, including: workflow arrangement: creating a new workflow in the computational workflow list interface and filling in basic configuration information, selecting and connecting different types of computing units in the workflow canvas interface to obtain a computational workflow, and using a drawer floating layer interaction form to configure the operation of each computing unit and submit the calculation; workflow execution: viewing the operation of the computational workflow in multiple aspects in the workflow canvas interface, including the task flow status between computing units, the operation status, operation records, operation monitoring and operation logs of each computing unit, and the operation monitoring of the workflow dimension; workflow result viewing: viewing the operation result file in the workflow canvas interface. This technical solution can provide users with a more convenient, efficient and intuitive workflow management tool, reduce computing costs, and support joint computing of computing units of multiple algorithms.

[0005] The above existing technologies all have the following problems: there are limitations in interface interaction; there are limitations in system performance and scalability; and there is a lack of error handling and recovery mechanisms. Summary of the invention

[0006] In view of the shortcomings of the prior art, the present invention proposes an intelligent workflow interactive configuration method and system based on a large model, which obtains the user's function realization demand information and combines multiple preset models, knowledge bases and plug-ins to build an overall large model. The large model can output corresponding output demand information based on the input demand information and support knowledge retrieval and configuration; the construction process involves steps such as model selection, knowledge base screening, plug-in integration and parameter adjustment; the output process includes the generation, evaluation, decision analysis and optimization selection of workflow configuration solutions. The present invention realizes the intelligence and automation of workflow configuration and improves configuration efficiency and accuracy.

[0007] To achieve the above object, the present invention provides the following technical solutions: The intelligent workflow interaction configuration method based on the big model includes: Acquire function realization requirement information input by the first user in the interactive interface, wherein the function realization requirement information includes input requirement information and corresponding output requirement information; According to the function realization requirement information, combined with multiple preset models, multi-knowledge base retrieval and multi-plug-in integration, an overall large model for realizing the function is constructed; Based on the overall large model, input demand information of a second user is input, and output demand information of the input demand information is output.

[0008] Specifically, the multiple models include a system reasoning model, an embedding model and a re-ranking model; the multiple knowledge bases include an industry standard knowledge base, domain knowledge, common sense knowledge, a business rule knowledge base, and a historical data knowledge base; the plug-ins include a data processing plug-in, a format conversion plug-in, and a visualization plug-in.

[0009] Specifically, the specific configuration of the knowledge base retrieval includes: A1: Determine the input requirement information and select the knowledge base to be searched. The knowledge base needs to be created in the system in advance. A2: performing a recall setting, wherein the recall setting is to search all knowledge bases at the same time, retrieve text fragments from multiple knowledge bases based on the query intention of the first user, and then select the answer that best meets the user's question from multiple search results through a re-ranking process to obtain a sorted and filtered search result list; A3: Based on the sorted and filtered search result list, according to business needs, the search results in the search result list are connected to the downstream processing nodes, and at the same time, the processing logic of the downstream nodes is configured. The downstream processing nodes include a question-and-answer system and a recommendation system.

[0010] Specifically, the construction process of the overall large model includes: B1: Preprocess the function realization requirement information, and conduct preliminary screening in multiple preset basic model libraries based on the preprocessed function realization requirement information to select the basic model; B2: Analyze the adaptability of the preliminarily selected basic models to the current functional implementation requirements, generate a candidate list of basic models that are adapted to the current functional implementation requirements, and attach an adaptability analysis report to each basic model in the candidate list; B3: Analyze the domain knowledge of the function realization requirement information, select the knowledge base resources that are highly matched with the determined domain from multiple preset knowledge bases, and extract and organize the selected knowledge base resources to obtain the domain knowledge graph that is closely related to the function requirement; B4: According to functional requirements, select corresponding plug-ins from multiple plug-ins, integrate the selected plug-ins, configure the interface between the plug-ins and the basic model, and generate a plug-in candidate list.

[0011] Specifically, the construction process of the overall large model also includes: B5: Based on the preprocessed function realization requirement information and the basic model candidate list, the domain knowledge graph closely related to the function requirement obtained in B3 is used in combination with the back propagation algorithm to adjust the parameters of the basic model to obtain a preliminarily adjusted basic model; B6: Use the plug-ins integrated in B4 to adjust and expand the architecture of the preliminarily adjusted basic model, reserve interfaces for data interaction and functional collaboration with the plug-ins, and generate a basic model instance with a plug-in integration interface; B7: Based on the basic model instance in B6 and the plug-in candidate list in B4, select the plug-in to be integrated, perform interface joint debugging, and generate a model system with integrated plug-ins; B8: The basic model instance, domain knowledge graph module and the model system with integrated plug-ins are finally assembled according to the business logic and functional flow to form an overall large model that realizes the functions, and the overall large model is verified.

[0012] Specifically, the output process of the overall large model includes: C1: Input the preprocessed workflow business requirement description information into the trained big model. The big model analyzes and infers the workflow based on the learned knowledge and patterns, and generates multiple preliminary workflow configuration solutions. C2: Evaluate multiple workflow configuration schemes generated by the large model according to a pre-set evaluation index system. At the same time, collect and organize the operation results of each workflow configuration scheme in the evaluation tool, and classify and summarize them according to the evaluation index system to form an evaluation report for each workflow configuration scheme. The evaluation index system includes efficiency index, cost index, quality index, flexibility index and compliance index. C3: Use multi-criteria decision analysis to assign weights to each evaluation indicator and calculate the comprehensive score of each configuration scheme; C4: Compare the comprehensive scores of all workflow configuration schemes and select the scheme with the highest score as the optimal workflow configuration scheme.

[0013] Specifically, the workflow configuration scheme includes workflow node definitions, connection relationships between nodes, task allocation rules, conditional judgment logic, and data transmission path information.

[0014] The intelligent workflow interactive configuration system based on the big model includes: demand acquisition module, big model construction module, knowledge configuration module, and workflow configuration module; The requirement acquisition module is used to acquire and pre-process the function realization requirement information input by the first user in the interactive interface; The large model building module is used to build an overall large model that realizes the function according to the preprocessed function realization requirement information, combined with multiple preset models, multiple knowledge bases and multiple plug-ins; The knowledge configuration module is used to support the knowledge retrieval requirements of the second user in the workflow and configure the processing logic of the downstream nodes according to the business requirements; The workflow configuration module is used to input the input requirement information of the second user based on the overall large model, and output the workflow configuration solution.

[0015] Specifically, the large model construction module includes: a model selection unit, a knowledge base screening unit, a plug-in integration unit, a parameter adjustment unit, and an assembly and verification unit; The model selection unit is used to perform preliminary screening in a preset plurality of basic model libraries according to the demand information and select an adapted basic model; The knowledge base screening unit is used to analyze the domain knowledge of the demand information, screen out knowledge base resources that are highly matched with it from multiple preset knowledge bases, and extract and organize the knowledge to form a domain knowledge graph; The plug-in integration unit is used to select a corresponding plug-in from multiple plug-ins according to functional requirements, integrate the plug-ins, and configure the interface between the plug-in and the basic model; The parameter adjustment unit is used to use the domain knowledge graph in combination with the back propagation algorithm to adjust the parameters of the basic model, and use the integrated plug-in to adjust and expand the model architecture; The assembly and verification unit is used to finally assemble the basic model instance, the domain knowledge graph module and the model system with integrated plug-ins according to the business logic and functional flow to form an overall large model and perform verification.

[0016] Specifically, the knowledge configuration module includes: a recall and sorting unit and a downstream node configuration unit; The recall and sorting unit is used to perform recall settings, search all knowledge bases, and re-sort the search results according to the query intent to select the answer that best meets the user's question; The downstream node configuration unit is used to connect the search results to the downstream processing nodes according to the sorted and filtered search result list and business requirements, and configure the processing logic of the downstream nodes.

[0017] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention proposes an intelligent workflow interactive configuration system based on a large model, and optimizes and improves the architecture, operation steps and processes. The system has the advantages of simple processes, low investment and operation costs, and low production work costs.

[0018] 2. The present invention proposes an intelligent workflow interaction configuration method based on a large model. By constructing an intelligent workflow interaction configuration method based on a large model, the intelligent and automated workflow configuration is realized. This method can quickly construct an overall large model based on user demand information and combine multiple models, knowledge bases and plug-ins, and output configuration information that meets user needs based on the model, thereby improving the efficiency of workflow configuration, reducing the tediousness and errors of manual configuration, and improving the accuracy and flexibility of configuration.

[0019] 3. The present invention proposes an intelligent workflow interactive configuration method based on a large model, and also ensures the quality and effect of the workflow configuration scheme through a detailed construction and output process. In the construction process, through the steps of preprocessing the functional requirement information, model selection, knowledge base screening and plug-in integration, a domain knowledge graph and a plug-in candidate list closely related to the functional requirements are formed. In the output process, multiple preliminary workflow configuration schemes are generated, evaluated and weighted, and finally the optimal configuration scheme is selected. This further ensures the rationality and feasibility of the workflow configuration scheme and improves the performance and efficiency of the entire workflow system. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 A schematic diagram of a large-model-based intelligent workflow interaction configuration method of the present invention; Figure 2 A schematic diagram of system variable settings for the intelligent workflow interaction configuration method based on a large model of the present invention; Figure 3 A schematic diagram of model configuration of a large-model-based intelligent workflow interaction configuration method of the present invention; Figure 4 This is a prompt word arrangement interface diagram of the intelligent workflow interaction configuration method based on a large model of the present invention; Figure 5 This is an architecture diagram of the intelligent workflow interaction configuration system based on a large model of the present invention. DETAILED DESCRIPTION

[0021] Example 1

[0022] See also Figure 1 , an embodiment provided by the present invention: a method for configuring intelligent workflow interaction based on a large model, comprising the following steps: Acquire function realization requirement information input by the first user in the interactive interface, wherein the function realization requirement information includes input requirement information and corresponding output requirement information; The first user uses a friendly interactive interface to describe the desired function in natural language in detail, including the required input data characteristics, format, source and other input requirement information, and the expected output result form, content, target and other output requirement information. For example, the first user may describe "I need a tool, the input is the employee attendance record form, and the output is the monthly attendance statistics report for each employee, including the number of attendance days, the number of late arrivals, the number of early departures and other information".

[0023] According to the function realization requirement information, combined with multiple preset models, multi-knowledge base retrieval and multi-plug-in integration, an overall large model for realizing the function is constructed; Based on the overall large model, input demand information of a second user is input, and output demand information of the input demand information is output.

[0024] To sum up, the overall implementation process can be summarized as: First, after the system is started, it enters the functional requirements acquisition stage. The system provides an intuitive and easy-to-use interactive interface to guide the first user to input the functional realization requirements information. The interface can use a variety of methods such as natural language input boxes, drop-down menus, sample templates, etc., to facilitate the first user to accurately describe his or her requirements. For example, through the natural language input box, the first user can specify in detail the type of input data, such as text files, database tables, images, etc., the format requirements of the data, such as CSV format, Excel format, and the specific form of the output results, such as reports, charts, text summaries, etc.; After obtaining the first user's function realization requirement information, the system enters the overall model construction stage. The system's built-in intelligent parsing module performs semantic analysis and understanding on the requirement information, extracting key requirement features and elements; then, the model selection module selects potentially applicable model components from multiple preset model libraries based on these features, such as data mining models, prediction models, classification models, etc.; at the same time, the knowledge extraction module searches and extracts knowledge and rules related to the requirements from multiple knowledge bases, such as industry specifications, business process knowledge, historical case data, etc.; the plug-in management module determines the plug-ins that need to be called according to the requirements, such as data cleaning plug-ins, format conversion plug-ins, encryption and decryption plug-ins, etc.; finally, the selected model components are combined and optimized through the model fusion engine, the relevant knowledge is integrated into the model using the knowledge injection mechanism, and the required plug-ins are seamlessly connected with the model through the plug-in integration framework to build a complete overall model that can realize the functions required by the first user; After the overall model is built, the system enters the actual application stage, waiting for the second user to input the demand information. When the second user inputs the relevant data or information according to the input method specified by the overall model, the execution engine of the overall model starts; the execution engine first pre-processes the input information, including data verification, format conversion, data cleaning and other operations to ensure the accuracy and availability of the input information; then, the pre-processed information is input into the constructed model, and the model performs calculations and analysis based on the internal algorithms and logic, uses the knowledge in the knowledge base to make decisions and judgments, and implements various specific functions through plug-ins, such as data visualization, file generation, and interaction with external systems; finally, the generated output demand information is presented to the second user in a predetermined manner, such as displaying a report on the screen, sending the result file by email, and storing the data in a specified database for subsequent query and use; During the implementation of the entire method, the system can also record and analyze the user's operations and feedback in order to continuously optimize the performance of the model and improve the accuracy and efficiency of the interactive configuration. For example, by analyzing the feedback information such as the second user's satisfaction with the output results, usage frequency, and modification suggestions, the system can automatically adjust the model's parameters, update the knowledge in the knowledge base, and optimize the plug-in's functions, so that the overall model can better meet the user's actual needs and achieve continuous improvement and optimization of the intelligent workflow interactive configuration.

[0025] The multiple models include a system reasoning model, an embedding model and a re-ranking model; the multiple knowledge bases include an industry standard knowledge base, domain knowledge, common sense knowledge, a business rule knowledge base, and a historical data knowledge base; the plug-ins include a data processing plug-in, a format conversion plug-in, and a visualization plug-in.

[0026] The specific configuration of knowledge base retrieval includes: A1: Determine the input requirement information and select the knowledge base to be searched. The knowledge base needs to be created in the system in advance. A2: performing a recall setting, wherein the recall setting is to search all knowledge bases at the same time, retrieve text fragments from multiple knowledge bases based on the query intention of the first user, and then select the answer that best meets the user's question from multiple search results through a re-ranking process to obtain a sorted and filtered search result list; A3: Based on the sorted and filtered search result list, according to business needs, the search results in the search result list are connected to the downstream processing nodes, and at the same time, the processing logic of the downstream nodes is configured. The downstream processing nodes include a question-and-answer system and a recommendation system.

[0027] Furthermore, the specific steps of connecting the search results to the downstream processing nodes include: (1) Determine the type, format, and quantity of search results, and clarify the type, function, and interface requirements of downstream processing nodes; (2) Ensure that the network connection between the search system and downstream processing nodes is normal, and prepare the necessary data transmission protocols and interface documents; (3) Sort search results according to business needs, such as by relevance, time, rating, etc., and filter search results according to preset filtering conditions, such as keyword matching, content quality, etc.; (4) According to the type and function of the downstream processing node, select the appropriate connection method, such as API call, message queue, and database sharing. According to the interface requirements of the downstream processing node, configure the corresponding interface parameters, such as URL, request header, and request body; (5) Send the sorted and filtered search results to the downstream processing node through the configured interface parameters; (6) The downstream processing node receives the search results from the search system and parses and processes them according to the data format and content of the search results; (7) Configure the processing logic of downstream processing nodes according to business needs, such as answer generation of question-answering systems and recommendation algorithms of recommendation systems. Then, the downstream processing nodes process the search results according to the configured processing logic and generate corresponding outputs. (8) Monitor the data transmission and processing between the search system and downstream processing nodes to ensure the integrity of data transmission and the accuracy of processing, and optimize and adjust the sorting, screening, connection method, processing logic, etc. of search results based on monitoring results and business needs.

[0028] For example, taking the question-and-answer system as an example, assuming that the search system has returned a set of search results about "artificial intelligence", these search results need to be connected to the downstream processing nodes of the question-and-answer system, and the processing logic of the question-and-answer system needs to be configured: (1) Sort the search results by relevance, and filter out the top 10 results most relevant to "artificial intelligence"; (2) Send the filtered search results to the interface of the question-and-answer system through API calls; (3) Configure the corresponding processing logic in the question-and-answer system, such as an answer generation algorithm based on natural language processing technology; (4) After receiving the search results, the question-and-answer system generates answers related to "artificial intelligence" according to the configured processing logic and returns them to the user. Through the above steps, the search results in the search result list can be effectively connected to the downstream processing nodes, and the corresponding business functions can be realized.

[0029] The construction process of the overall large model includes: B1: Preprocess the function realization requirement information, and conduct preliminary screening in multiple preset basic model libraries based on the preprocessed function realization requirement information to select the basic model; B2: Analyze the adaptability of the preliminarily selected basic models to the current functional implementation requirements, generate a candidate list of basic models that are adapted to the current functional implementation requirements, and attach an adaptability analysis report to each basic model in the candidate list; Furthermore, the specific steps of B2 include: (1) Clarify functional requirements: Understand the specific requirements for the current functional implementation, including performance indicators, input and output requirements, and data processing capabilities; (2) Basic model evaluation: Evaluate the initially selected basic models one by one, and analyze their technical characteristics, performance, and scope of application. During the evaluation process, reference can be made to existing test data, user feedback, and expert opinions. (3) Adaptability analysis: Analyze the adaptability of each basic model to the functional requirements; (4) Candidate list generation: Based on the results of the adaptability analysis, the basic models with higher adaptability are included in the candidate list. The candidate list can be sorted from high to low according to the adaptability for subsequent selection; (5) Verification and testing: Further verification and testing of the models in the candidate list to ensure that their performance in actual applications meets expectations. Testing may include unit testing, integration testing, and performance testing. (6) Final selection: Based on the test results, select the basic model that best meets the current functional implementation requirements.

[0030] B3: Analyze the domain knowledge of the function realization requirement information, select the knowledge base resources that are highly matched with the determined domain from multiple preset knowledge bases, and extract and organize the selected knowledge base resources to obtain the domain knowledge graph that is closely related to the function requirement; Furthermore, the specific steps of B3 include: (1) Functional requirements analysis and domain determination: clarify the specific domain knowledge required for function implementation and determine the core concepts and scope of the domain knowledge; (2) Knowledge base resource screening: From multiple preset knowledge bases, screen according to the core concepts and scope of domain knowledge, evaluate the matching degree between knowledge base resources and functional requirements, and select knowledge base resources with high matching degree; (3) Knowledge extraction: Using rule-based knowledge extraction methods to extract knowledge from the screened knowledge base resources, including entity extraction, relationship extraction, and attribute extraction. The rule-based knowledge extraction method is the prior art content in this field and is not an inventive solution of this application, so it will not be described in detail here; (4) Knowledge organization and graph construction: Organize the extracted knowledge to form a structured knowledge representation, and build a domain knowledge graph based on the knowledge representation, including the definition and connection of elements such as nodes, relationships, and attributes; (5) Knowledge graph verification and optimization: Verify the constructed knowledge graph to ensure the accuracy and completeness of the knowledge, and optimize and adjust the knowledge graph based on the verification results.

[0031] B4: Select corresponding plug-ins from multiple plug-ins according to functional requirements, integrate the selected plug-ins, configure the interface between the plug-ins and the basic model, and generate a plug-in candidate list; B5: Based on the preprocessed function realization requirement information and the basic model candidate list, the domain knowledge graph closely related to the function requirement obtained in B3 is used in combination with the back propagation algorithm to adjust the parameters of the basic model to obtain a preliminarily adjusted basic model; Furthermore, the specific steps of B5 include: (1) Data preprocessing: Clean, standardize, and normalize the function realization requirement information to ensure data quality and consistency, and convert the processed data into a format acceptable to the model; (2) Constructing a domain knowledge graph: Using the domain knowledge closely related to functional requirements obtained in B3, construct or improve the knowledge graph to ensure that the entities, relationships, and attributes in the knowledge graph are relevant to functional requirements; (3) Select the basic model: From the basic model candidate list, select the most appropriate model as the starting point based on the functional requirements and the characteristics of the domain knowledge graph; (4) Initializing model parameters: performing random initialization on the parameters of the selected basic model, wherein the random initialization of parameters is a prior art in the art and is not an inventive solution of the present application, and is not described in detail herein; (5) Combine knowledge graph for forward propagation: Input the preprocessed data into the model, and use the information in the domain knowledge graph to assist the model in forward propagation calculation; (6) Calculating a loss function: Calculating a loss function based on the output of the model and the true label, wherein the loss function calculation formula is the prior art content in the art and is not an inventive solution of the present application, and is not described in detail here; (7) Back propagation and parameter update: using the back propagation algorithm, the gradient is calculated layer by layer from the output layer to the input layer, and the SGD optimization algorithm is used to update the model parameters according to the gradient calculation results. The back propagation algorithm and the SGD optimization algorithm are the prior art contents in this field, and are not the inventive solutions of this application, and are not described in detail here; (8) Iterative training: Repeat steps (5) to (7) until the performance of the model on the validation set reaches the preset stopping condition; (9) Obtaining a preliminary adjusted basic model: After multiple iterations of training, a preliminary adjusted basic model is obtained.

[0032] B6: Use the plug-ins integrated in B4 to adjust and expand the architecture of the preliminarily adjusted basic model, reserve interfaces for data interaction and functional collaboration with the plug-ins, and generate a basic model instance with a plug-in integration interface; B7: According to the basic model instance in B6 and the plug-in candidate list in B4, the plug-in to be integrated is selected, and the interface joint debugging is performed to generate a model system with integrated plug-ins. The interface joint debugging is the prior art content in this field and is not the inventive solution of this application, and will not be described in detail here; B8: The basic model instance, domain knowledge graph module and the model system with integrated plug-ins are finally assembled according to the business logic and functional flow to form an overall large model that realizes the functions, and the overall large model is verified.

[0033] Furthermore, the specific steps of B8 include: (1) Component preparation: Ensure that the basic model instance has been trained and achieves the expected performance, the domain knowledge graph module has been built and contains entities, relationships, and attributes that are closely related to functional requirements, and that the model system with integrated plug-ins has been developed and preliminarily tested; (2) Design the architecture of the overall large model based on business logic and functional processes, and determine the interfaces and data interaction methods between components; (3) Component integration: Integrate the basic model instance, domain knowledge graph module, and the model system with integrated plug-ins according to the designed architecture to ensure smooth data interaction and interface matching between components; (4) Functional testing: Perform functional testing on the overall large model to verify whether it meets the functional requirements; (5) Record problems found during the test and fix them. Based on the test results, optimize the performance of the overall large model.

[0034] Example 2

[0035] See also Figure 2-Figure 4 In this embodiment, the output process of the overall large model includes: C1: Input the preprocessed workflow business requirement description information into the trained big model. The big model analyzes and infers the workflow based on the learned knowledge and patterns, and generates multiple preliminary workflow configuration solutions. The workflow simplifies the complexity of the system by breaking down complex tasks into a series of smaller, manageable steps. This approach reduces the reliance on prompt words and model reasoning capabilities, enhances performance when processing complex tasks, and improves the interpretability, stability, and fault tolerance of the system. Specifically, it includes: (1) Creating and starting a workflow: To start a workflow, you can choose to create a new workflow from scratch. In this process, you need to master some basic operating skills, such as how to add nodes in the workspace, how to connect nodes to each other, how to configure nodes, and how to debug the workflow and view its running records; after completing the construction of the workflow, save and publish the workflow so that it can be used; finally, execute the workflow through the published application.

[0036] (2) Variables: Workflow applications are composed of multiple independent nodes. Most nodes have input and output items, but the input information and output results required by each node are different. In order to use a fixed symbol to represent dynamically changing content, variables are used as a dynamic data container to store and transmit non-fixed information and reference each other between different nodes, thereby achieving flexible communication. (3) Nodes: Nodes are the basic elements for building workflows. By connecting nodes with different functions, continuous operations in the workflow can be completed, including: 1) In each workflow application, there is a preset node named "Start". The function of this node is to provide the necessary initial information and data for the start of the entire workflow, such as user input information, to ensure that the workflow can proceed smoothly. At the same time, on the settings page of the Start node, you can see two parts of settings, namely input fields and preset system variables, such as Figure 2 As shown; 2) Use LLM nodes to process the information entered by users in the nodes, and use its multiple capabilities such as dialogue, generation, classification and processing in the workflow to efficiently handle various tasks. It can also be applied in different links of the workflow. The configuration steps of LLM include: a. In the application editing page, you can add an LLM node by right-clicking the mouse, dragging the + sign at the end of the previous node, or dragging the LLM on the left; b . Configure model parameters, which will affect the output of the model, such as temperature, TopP, maximum number of tags, response format, etc. The system provides three sets of preset parameters: creativity, balance and precision. If you are not familiar with these parameters, you can choose the default settings, such as Figure 3 As shown; c. Write prompt words, including system and user parts, such as Figure 4 As shown in the figure, in the prompt word editor, you can enter " / " or click "(x)" to call up the variable insertion menu and insert the upstream node variable into the prompt word as context content; 3) Retrieve text content related to the user’s question from the knowledge base.

[0037] C2: Evaluate multiple workflow configuration schemes generated by the large model according to a pre-set evaluation index system. At the same time, collect and organize the operation results of each workflow configuration scheme in the evaluation tool, and classify and summarize them according to the evaluation index system to form an evaluation report for each workflow configuration scheme. The evaluation index system includes but is not limited to efficiency indicators, cost indicators, quality indicators, flexibility indicators and compliance indicators; C3: Use multi-criteria decision analysis methods Assign weights to each evaluation indicator and calculate the comprehensive score of each configuration solution , where A represents the judgment matrix constructed after comparing and scoring the evaluation indicators pairwise. represents the maximum eigenvalue of the judgment matrix, Indicates j The comprehensive score of the workflow configuration scheme, w represents the weight, m represents the number of evaluation indicators, Indicates i The weights assigned to the evaluation indicators are Indicates j The workflow configuration scheme is in i Quantitative scores on evaluation indicators; It should be noted that A is a The matrix of is a scalar, w is a Vector.

[0038] C4: Compare the comprehensive scores of all workflow configuration schemes and select the scheme with the highest score as the optimal workflow configuration scheme.

[0039] The workflow configuration scheme includes the workflow node definition, the connection relationship between nodes, task allocation rules, condition judgment logic and data transmission path information.

[0040] Example 2

[0041] See also Figure 5 Another embodiment provided by the present invention is a large-model-based intelligent workflow interactive configuration system, comprising: Requirements acquisition module, large model building module, knowledge configuration module, workflow configuration module; A requirement acquisition module, used to acquire and pre-process the function realization requirement information input by the first user in the interactive interface, so as to provide a basis for the subsequent large model construction; A large model building module is used to build an overall large model that realizes the function according to the preprocessed function realization requirement information, combined with multiple preset models, multiple knowledge bases and multiple plug-ins; The knowledge configuration module is used to support the knowledge retrieval needs of the second user in the workflow and configure the processing logic of the downstream nodes according to business needs; The workflow configuration module is used to input the input requirement information of the second user based on the overall large model, and output the workflow configuration plan, while supporting the evaluation and optimization of the configuration plan.

[0042] The large model building module includes: model selection unit, knowledge base screening unit, plug-in integration unit, parameter adjustment unit, assembly and verification unit; A model selection unit is used to perform preliminary screening in a preset multiple basic model library according to the demand information and select an adapted basic model; The knowledge base screening unit is used to analyze the domain knowledge of the demand information, screen out knowledge base resources that are highly matched with it from multiple preset knowledge bases, and extract and organize the knowledge to form a domain knowledge graph; The plug-in integration unit is used to select corresponding plug-ins from multiple plug-ins according to functional requirements, integrate them, and configure the interface between the plug-in and the basic model; The parameter adjustment unit is used to use the domain knowledge graph and the back propagation algorithm to adjust the parameters of the basic model, and use the integrated plug-ins to adjust and expand the model architecture; The assembly and verification unit is used to assemble the basic model instance, domain knowledge graph module and the model system with integrated plug-ins according to the business logic and functional flow to form an overall large model and perform verification.

[0043] The knowledge configuration module includes: a recall and sorting unit and a downstream node configuration unit; The recall and ranking unit is used to perform recall settings, search all knowledge bases, and re-rank search results according to the query intent to select the answer that best matches the user's question; The downstream node configuration unit is used to connect the search results to the downstream processing nodes according to the sorted and filtered search result list and business requirements, and configure the processing logic of the downstream nodes.

[0044] The workflow configuration module includes: workflow generation unit, evaluation unit, decision analysis unit, and optimization selection unit; A workflow generation unit is used to input the pre-processed workflow business requirement description information into the trained large model to generate multiple preliminary workflow configuration schemes; An evaluation unit, used to evaluate multiple workflow configuration schemes generated by the large model according to a pre-set evaluation index system and form an evaluation report; A decision analysis unit is used to assign weights to various evaluation indicators using a multi-criteria decision analysis method and calculate a comprehensive score for each configuration scheme; The optimization selection unit is used to compare the comprehensive scores of all workflow configuration schemes and select the scheme with the highest score as the optimal workflow configuration scheme.

[0045] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation modes, which are merely illustrative rather than restrictive. Under the guidance of the present invention, ordinary technicians in the field may also change, modify, replace and modify the above-mentioned embodiments without departing from the purpose and scope of protection of the present invention, and all of these are within the protection of the present invention.

[0046] If the disclosed technical solution involves personal information, the product using the disclosed technical solution has clearly informed the personal information processing rules and obtained the individual's voluntary consent before processing the personal information. If the disclosed technical solution involves sensitive personal information, the product using the disclosed technical solution has obtained the individual's separate consent before processing the sensitive personal information, and at the same time meets the "explicit consent" requirement. For example, on personal information collection devices such as cameras, clear and prominent signs are set to inform that the personal information collection scope has been entered and personal information will be collected. If the individual voluntarily enters the collection scope, it is deemed that he or she agrees to the collection of his or her personal information; or on the device that processes personal information, when the personal information processing rules are notified by obvious signs / information, the individual's authorization is obtained through pop-up information or by asking the individual to upload his or her personal information; among them, the personal information processing rules may include information such as the personal information processor, the purpose of personal information processing, the processing method, and the type of personal information processed.

Claims

1. An intelligent workflow interaction configuration method based on a large model, characterized in that: include: Acquire function realization requirement information input by the first user in the interactive interface, wherein the function realization requirement information includes input requirement information and corresponding output requirement information; According to the function realization requirement information, combined with multiple preset models, multi-knowledge base retrieval and multi-plug-in integration, an overall large model for realizing the function is constructed; Based on the overall large model, input demand information of a second user is input, and output demand information of the input demand information is output.

2. The method for configuring intelligent workflow interaction based on a large model as claimed in claim 1, characterized in that: The multiple models include a system reasoning model, an embedding model and a re-ranking model; the multiple knowledge bases include an industry standard knowledge base, domain knowledge, common sense knowledge, a business rule knowledge base, and a historical data knowledge base; the plug-ins include a data processing plug-in, a format conversion plug-in, and a visualization plug-in.

3. The method for configuring intelligent workflow interaction based on a large model as claimed in claim 2, characterized in that: The specific configuration of the knowledge base retrieval includes: A1: Determine the input requirement information and select the knowledge base to be searched. The knowledge base needs to be created in the system in advance. A2: performing a recall setting, wherein the recall setting is to search all knowledge bases at the same time, retrieve text fragments from multiple knowledge bases based on the query intention of the first user, and then select the answer that best meets the user's question from multiple search results through a re-ranking process to obtain a sorted and filtered search result list; A3: Based on the sorted and filtered search result list, according to business needs, the search results in the search result list are connected to the downstream processing nodes, and at the same time, the processing logic of the downstream nodes is configured. The downstream processing nodes include a question-and-answer system and a recommendation system.

4. The method for configuring intelligent workflow interaction based on a large model as claimed in claim 3, characterized in that: The construction process of the overall large model includes: B1: Preprocess the function realization requirement information, and conduct preliminary screening in multiple preset basic model libraries based on the preprocessed function realization requirement information to select the basic model; B2: Analyze the adaptability of the preliminarily selected basic models to the current functional implementation requirements, generate a candidate list of basic models that are adapted to the current functional implementation requirements, and attach an adaptability analysis report to each basic model in the candidate list; B3: Analyze the domain knowledge of the function realization requirement information, select the knowledge base resources that are highly matched with the determined domain from multiple preset knowledge bases, and extract and organize the selected knowledge base resources to obtain the domain knowledge graph that is closely related to the function requirement; B4: According to functional requirements, select corresponding plug-ins from multiple plug-ins, integrate the selected plug-ins, configure the interface between the plug-ins and the basic model, and generate a plug-in candidate list.

5. The method for configuring intelligent workflow interaction based on a large model as claimed in claim 4, characterized in that: The construction process of the overall large model also includes: B5: Based on the preprocessed function realization requirement information and the basic model candidate list, the domain knowledge graph closely related to the function requirement obtained in B3 is used in combination with the back propagation algorithm to adjust the parameters of the basic model to obtain a preliminarily adjusted basic model; B6: Use the plug-ins integrated in B4 to adjust and expand the architecture of the preliminarily adjusted basic model, reserve interfaces for data interaction and functional collaboration with the plug-ins, and generate a basic model instance with a plug-in integration interface; B7: Based on the basic model instance in B6 and the plug-in candidate list in B4, select the plug-in to be integrated, perform interface joint debugging, and generate a model system with integrated plug-ins; B8: The basic model instance, domain knowledge graph module and the model system with integrated plug-ins are finally assembled according to the business logic and functional flow to form an overall large model that realizes the functions, and the overall large model is verified.

6. The method for configuring intelligent workflow interaction based on a large model as claimed in claim 5, characterized in that: The output process of the overall large model includes: C1: Input the preprocessed workflow business requirement description information into the trained big model. The big model analyzes and infers the workflow based on the learned knowledge and patterns, and generates multiple preliminary workflow configuration solutions. C2: Evaluate multiple workflow configuration schemes generated by the large model according to a pre-set evaluation index system. At the same time, collect and organize the operation results of each workflow configuration scheme in the evaluation tool, and classify and summarize them according to the evaluation index system to form an evaluation report for each workflow configuration scheme. The evaluation index system includes efficiency index, cost index, quality index, flexibility index and compliance index. C3: Use multi-criteria decision analysis to assign weights to each evaluation indicator and calculate the comprehensive score of each configuration scheme; C4: Compare the comprehensive scores of all workflow configuration schemes and select the scheme with the highest score as the optimal workflow configuration scheme.

7. The method for configuring intelligent workflow interaction based on a large model as claimed in claim 6, characterized in that: The workflow configuration scheme includes workflow node definitions, connection relationships between nodes, task allocation rules, conditional judgment logic, and data transmission path information.

8. A large-model-based intelligent workflow interactive configuration system, which is used to implement the large-model-based intelligent workflow interactive configuration method described in any one of claims 1 to 7, characterized in that: include: Requirements acquisition module, large model building module, knowledge configuration module, workflow configuration module; The requirement acquisition module is used to acquire and pre-process the function realization requirement information input by the first user in the interactive interface; The large model building module is used to build an overall large model that realizes the function according to the preprocessed function realization requirement information, combined with multiple preset models, multiple knowledge bases and multiple plug-ins; The knowledge configuration module is used to support the knowledge retrieval requirements of the second user in the workflow and configure the processing logic of the downstream nodes according to the business requirements; The workflow configuration module is used to input the input requirement information of the second user based on the overall large model, and output the workflow configuration solution.

9. The large model-based intelligent workflow interactive configuration system according to claim 8, characterized in that: The large model construction module includes: a model selection unit, a knowledge base screening unit, a plug-in integration unit, a parameter adjustment unit, and an assembly and verification unit; The model selection unit is used to perform preliminary screening in a preset plurality of basic model libraries according to the demand information and select an adapted basic model; The knowledge base screening unit is used to analyze the domain knowledge of the demand information, screen out knowledge base resources that are highly matched with it from multiple preset knowledge bases, and extract and organize the knowledge to form a domain knowledge graph; The plug-in integration unit is used to select a corresponding plug-in from multiple plug-ins according to functional requirements, integrate the plug-ins, and configure the interface between the plug-in and the basic model; The parameter adjustment unit is used to use the domain knowledge graph in combination with the back propagation algorithm to adjust the parameters of the basic model, and use the integrated plug-in to adjust and expand the model architecture; The assembly and verification unit is used to finally assemble the basic model instance, the domain knowledge graph module and the model system with integrated plug-ins according to the business logic and functional flow to form an overall large model and perform verification.

10. The large model-based intelligent workflow interactive configuration system according to claim 9, characterized in that: The knowledge configuration module includes: a recall and sorting unit and a downstream node configuration unit; The recall and sorting unit is used to perform recall settings, search all knowledge bases, and re-sort the search results according to the query intent to select the answer that best meets the user's question; The downstream node configuration unit is used to connect the search results to the downstream processing nodes according to the sorted and filtered search result list and business requirements, and configure the processing logic of the downstream nodes.

Images