Interactive information processing method, electronic device, storage medium, and dialogue system

Abstract

The embodiments of the present disclosure relate to the fields of large-model technology and information processing. Provided are an interactive information processing method, an electronic device, a storage medium, and a dialogue system. The method comprises: acquiring input information to be replied to; acquiring configuration information associated with the input information, wherein the configuration information is used for describing an initial information generation rule that an information processing model needs to satisfy during the process of generating reply information corresponding to the input information, and the initial information generation rule is used for guiding the information processing model to determine a target information generation rule that the information processing model needs to satisfy during the process of generating the reply information; on the basis of the configuration information and pre-acquired parameter information associated with the input information, generating prompt information of the information processing model, wherein the parameter information is used for defining key content in the reply information; and using the prompt information to generate the reply information comprising the key content.

Description

[0001] The processing method of interactive information, electronic equipment, storage medium and dialogue system cross reference The present disclosure claims priority to the Chinese patent application No. 202410850812.9, filed on June 27, 2024, entitled "Processing method of interactive information, electronic equipment, storage medium and dialogue system", the entire content of which is incorporated herein by reference. TECHNICAL FIELD The present disclosure relates to the field of large model technology and information processing, in particular, to a processing method of interactive information, an electronic equipment, a storage medium and a dialogue system. BACKGROUND Currently, with the continuous development of artificial intelligence technology, dialogue robots are increasingly widely used in many fields. Natural language processing (NLP) technology plays a crucial role in the development of dialogue robots. Therefore, it is very important to use dialogue robots to analyze user input information and provide corresponding reply information to realize dialogue with the user. In related technologies, the dialogue system in the dialogue robot can be set up in a pipeline architecture. The architecture includes different components, each component is responsible for different tasks to achieve the effect of processing user input information and reply generation. For example, the architecture can include a natural language understanding (NLU) component, a dialogue management (DM) component and a natural language generation (NLG) component. Through the above component setting, the operator can directly participate in the definition and description of the task flow of the dialogue, that is, the above dialogue system relies too much on the configuration of the operator. For example, the operator needs to configure a large amount of intent data for NLU. However, it is inevitable to have intent conflicts and other situations during the configuration process, which greatly affects the intent recognition effect of NLU. Through the above analysis, it can be understood that due to the high participation of the operator in the task of the entire dialogue system, the management flexibility of the entire dialogue system is poor, and there is also a problem of accumulation of human errors. Therefore, there is still a technical problem of low processing efficiency of interactive information. For the above problems, no effective solution has been proposed so far. SUMMARY According to an aspect of the present disclosure, a processing method of interactive information is provided. The method can include:Obtaining input information to be replied; obtaining configuration information associated with the input information, wherein the configuration information is used to describe an initial information generation rule required to be met in a process in which an information processing model generates reply information corresponding to the input information, and the initial information generation rule is used to guide the information processing model to determine a target information generation rule required to be met in the process of generating the reply information; generating prompt information of the information processing model based on the configuration information and pre-obtained parameter information associated with the input information, wherein the parameter information is used to define key content in the reply information; and generating the reply information including the key content by using the prompt information. According to another aspect of the embodiments of the present disclosure, a processing method of interactive information is provided. The method can include: obtaining input information to be replied; determining a scene to which the input information belongs; obtaining configuration information associated with the input information in the scene, wherein the configuration information is used to describe an initial information generation rule required to be met in a process in which an information processing model generates reply information corresponding to the input information, and the initial information generation rule is used to guide the information processing model to determine a target information generation rule required to be met in the process of generating the reply information; generating prompt information of the information processing model based on the configuration information and pre-obtained parameter information associated with the input information, wherein the parameter information is used to define key content in the reply information; and generating the reply information satisfying the scene and including the key content by using the prompt information. According to another aspect of the embodiments of the present disclosure, a processing method of interactive information is provided. The method can include: determining an initial information generation rule required to be met in a process in which an information processing model generates reply information corresponding to input information, wherein the initial information generation rule is used to guide the information processing model to determine a target information generation rule required to be met in the process of generating the reply information; and creating configuration information based on the initial information generation rule, wherein the configuration information is used to generate prompt information of the information processing model, and the prompt information is used to generate the reply information. According to another aspect of the embodiments of the present disclosure, a processing method of interactive information is provided. The method can include: obtaining inquiry information, wherein the inquiry information is used to represent information to be replied in a task-oriented dialogue scene; obtaining configuration information associated with the inquiry information, wherein the configuration information is used to describe an initial information generation rule required to be met in a process in which a dialogue model generates reply information corresponding to the inquiry information, and the initial information generation rule is used to guide the dialogue model to determine a target information generation rule required to be met in the process of generating the reply information, and the dialogue model is trained based on inquiry information samples in the task-oriented dialogue scene and reply information samples in the task-oriented dialogue scene; generating prompt information of the dialogue model based on the configuration information and pre-obtained parameter information associated with the inquiry information, wherein the parameter information is used to define key content in the reply information; and generating the reply information satisfying the task-oriented dialogue scene and including the key content by using the prompt information.The parameter information is used to define the key content in the reply information. The prompt information is used to guide the dialogue model to generate the reply information including the key content according to the initial information generation rule and the target information generation rule. According to another aspect of the embodiments of the present disclosure, a dialogue system is provided. The dialogue system can include: a client configured to obtain input information to be replied; and a server configured to send configuration information associated with the input information to the client, wherein the configuration information is used to describe an initial information generation rule required to be met in a process in which an information processing model generates reply information corresponding to the input information, and the initial information generation rule is used to guide the information processing model to determine a target information generation rule required to be met in the process of generating the reply information; and wherein the client is configured to generate prompt information of the information processing model based on the configuration information and pre-acquired parameter information associated with the input information, and generate the reply information including the key content by using the prompt information, and the parameter information is used to define the key content in the reply information. According to another aspect of the embodiments of the present disclosure, an electronic device is also provided, which can include a memory and a processor: the memory is configured to store computer executable instructions, and the processor is configured to execute the computer executable instructions. When the computer executable instructions are executed by the processor, the processing method of the interactive information of any one of the above is implemented. According to another aspect of the embodiments of the present disclosure, a processor is also provided, which is configured to run a program, wherein the processing method of the interactive information of any one of the above is executed when the program is running. According to another aspect of the embodiments of the present disclosure, a computer readable storage medium is also provided, which includes a stored program, wherein the processing method of the interactive information of any one of the above is executed when the program is running. According to another aspect of the embodiments of the present disclosure, a computer program product is also provided, which includes a non-volatile computer readable storage medium, and the non-volatile computer readable storage medium stores a computer program, and the computer program is executed by the processor to implement the processing method of the interactive information of any one of the above. According to another aspect of the embodiments of the present disclosure, a computer program product is also provided, which includes a computer program, and the computer program is executed by the processor to implement the processing method of the interactive information of any one of the above. According to another aspect of the embodiments of the present disclosure, a computer program product is also provided, which includes a non-volatile computer readable storage medium, and the non-volatile computer readable storage medium stores a computer program, and the computer program is executed by the processor to implement the processing method of the interactive information of any one of the above. According to another aspect of the embodiments of the present disclosure, a computer program is also provided, and the computer program is executed by the processor to implement the processing method of the interactive information of any one of the above. In the embodiments of the present disclosure,If a response is required to the input information to be answered, the aforementioned input information can be obtained and fed into the information processing model for processing. During processing, the initial information generation rules that the input information must satisfy can be determined, i.e., the configuration information of the information processing model. Based on the configuration information and the pre-acquired corresponding parameter information, prompt information can be generated to guide the information processing model in processing the input information and generating response information. Through the prompt information, the information processing model is guided to determine the target information generation rules that must be satisfied during the process of generating response information. Thus, response information containing key content can be generated according to the target information generation rules. Since excessive involvement of operators in the configuration of the dialogue system can lead to a decrease in the efficiency of the entire interactive information processing, this embodiment of the disclosure can define configuration information related to the task flow of interactive information processing in the dialogue system. This configuration information is easy to use and maintain, and can effectively avoid providing a large amount of interfering information to the information processing model. Based on this, the embodiments of this disclosure rely on the reasoning ability and instruction compliance of the information processing model. By utilizing the prompt information generated from the configuration information, the information processing model can determine the target information generation rules during the generation of response information. That is, it can autonomously plan the process without requiring manual definition of the corresponding dialogue logic. This achieves the goal of enabling the information processing model to smoothly and naturally generate response information to the input information, thereby improving the technical efficiency of interactive information processing and solving the technical problem of low processing efficiency of interactive information. It is worth noting that the above general description and the following detailed description are merely illustrative and explanatory and do not constitute a limitation of this disclosure. The accompanying drawings, which are included to provide a further understanding of this disclosure and constitute a part of this disclosure, illustrate exemplary embodiments of this disclosure and are used to explain this disclosure, but do not constitute an improper limitation of this disclosure. In the accompanying drawings: Figure 1 is a schematic diagram of an application scenario of an interactive information processing method according to an embodiment of the present disclosure; Figure 2 is a flowchart of an interactive information processing method according to an embodiment of the present disclosure; Figure 3 is a flowchart of another interactive information processing method according to an embodiment of the present disclosure; Figure 4 is a flowchart of another interactive information processing method according to an embodiment of the present disclosure; Figure 5 is a flowchart of another interactive information processing method according to an embodiment of the present disclosure; Figure 6 is a schematic diagram of a dialogue system according to an embodiment of the present disclosure; Figure 7 is a schematic diagram of a dialogue system with a pipeline architecture in the related art; Figure 8 is a schematic diagram of the process of generating response results in a dialogue system based on a pipeline architecture in the related art;9 is a schematic diagram of a logical architecture of an end-to-end large model task-oriented dialogue according to an embodiment of the present disclosure; FIG. 10 is a flowchart of processing of an end-to-end large model task-oriented dialogue according to an embodiment of the present disclosure; FIG. 11 is a schematic diagram of a flowchart of an electronic toll collection system transaction according to an embodiment of the present disclosure; FIG. 12 is a schematic diagram of calling an external application programming interface plug-in according to an embodiment of the present disclosure; FIG. 13 is a schematic diagram of a complex instruction in a high-level mode according to an embodiment of the present disclosure; FIG. 14 is a schematic diagram of an application programming interface plug-in of an electronic toll collection system transaction according to an embodiment of the present disclosure; FIG. 15 is a schematic diagram of a parameter of an electronic toll collection system transaction according to an embodiment of the present disclosure; FIG. 16 is a schematic diagram of a local high-speed information query system flowchart in a related art; FIG. 17 is a schematic diagram of an end-to-end flowchart of a high-speed information query system according to an embodiment of the present disclosure; FIG. 18 is a schematic diagram of a dialogue question and answer effect in a high-speed information query system according to an embodiment of the present disclosure; FIG. 19 is a schematic diagram of a reply content in a high-speed information query system according to an embodiment of the present disclosure; FIG. 20 is a schematic diagram of an end-to-end flowchart of a photo developing task system according to an embodiment of the present disclosure; FIG. 21 is a schematic diagram of a dialogue question and answer effect of a photo developing task system according to an embodiment of the present disclosure; FIG. 22 is a schematic diagram of a processing device of interactive information according to an embodiment of the present disclosure; FIG. 23 is a schematic diagram of another processing device of interactive information according to an embodiment of the present disclosure; FIG. 24 is a schematic diagram of another processing device of interactive information according to an embodiment of the present disclosure; FIG. 25 is a schematic diagram of another processing device of interactive information according to an embodiment of the present disclosure; FIG. 26 is a structural block diagram of a computer terminal according to an embodiment of the present disclosure; FIG. 27 is a block diagram of an electronic device of a processing method of interactive information according to an embodiment of the present disclosure; FIG. 28 is a hardware structural block diagram of a computer terminal (or mobile device) for implementing a processing method of interactive information according to an embodiment of the present disclosure; FIG. 29 is a structural block diagram of a computing environment of a processing method of interactive information according to an embodiment of the present disclosure. DETAILED DESCRIPTION In order for those skilled in the art to better understand the present disclosure, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below with reference to the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are only a part of the embodiments of the present disclosure, but not all the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by those skilled in the art without creative labor should fall within the scope of protection of the present disclosure. It should be noted that the terms "first", "second" in the specification and claims of the present disclosure and the above-mentioned drawingsand their derivatives are intended to cover open, non- exclusive inclusion, such that a process, method, system, product or apparatus that comprises an "includes" or "has" does not, without additional structure, require the existence of every enumerated step, component, feature, or element. The technical solutions provided by the present disclosure are mainly implemented by using large model technology. Here, the large model refers to a deep learning model with large-scale model parameters, which usually contains hundreds of millions, tens of billions, hundreds of billions, thousands of billions, or even tens of thousands of billions of model parameters. The large model can also be referred to as a foundation model. A large-scale unlabeled corpus is used to pre-train the large model, and a pre-training model with more than one hundred million parameters is output. The model can adapt to a wide range of downstream tasks and has good generalization ability. For example, a large-scale language model (Language Model for Information Retrieval, LLM) and a multi-modal pre-training model. It should be noted that the large model can be fine-tuned by a small amount of samples in actual application, so that the large model can be applied to different tasks. For example, the large model can be widely used in the fields of NLP, computer vision, and speech processing. Specifically, the large model can be applied to computer vision field tasks such as visual question answering (Visual Question Answering, VQA), image captioning (Image Captioning, IC), and image generation. The large model can also be widely used in natural language processing field tasks such as text-based sentiment classification, text summary generation, and machine translation. Therefore, the main application scenarios of the large model include but are not limited to digital assistants, intelligent robots, search, online education, office software, e-commerce, intelligent design, and the like. In the embodiments of the present disclosure, the application of the generated program by using the target language processing model trained by the generative large model in the dialogue scenario is taken as an example for explanation and description. First, some nouns or terms appearing in the process of describing the embodiments of the present disclosure are applicable to the following explanations: a large language model (Large Language Model, LLM) is a deep learning model with large-scale parameters and complexity.Machine learning models with complex structures are often used to handle natural language-related tasks. Training these models requires a large amount of data and computing resources, and they can effectively handle complex natural language processing tasks. Agents are intelligent entities based on large model technology that help users complete complex tasks. Internally, they generally contain task flows and related application programming interfaces (APIs). These agents are commonly used in virtual assistants, intelligent customer service, and automated task processing, among other fields. They use large-scale artificial intelligence models and related task flows to achieve task-oriented intelligent dialogue and operations. Prompts are texts or instructions input to large models to guide them to generate more accurate and targeted output results. When using large models for natural language processing tasks, prompts can serve as a guiding input to help the model produce specific text generation or responses. The setting of prompts can significantly influence the model's output, helping it better understand user intent and generate relevant content. In the task-oriented dialogue method based on large model agents, flows are components used to describe specific task logic. Users can arrange task logic through a canvas, and agents complete complex tasks based on the defined task logic arrangement. This flow setting helps agents better understand user needs and perform corresponding operations.

[0002] API Plugin, is a component used to define an external API service that can be flexibly invoked by the Agent, thereby realizing the integration and interaction of the Agent with external systems. The setting of the above API Plugin can make the Agent have more functions and flexibility, and be able to call the services and data provided by external systems when processing task-oriented dialogues. Through the API Plugin, the Agent can communicate with external systems to obtain the data or services provided by the external systems. The above external systems can include third-party service providers, internal systems of enterprises or other cloud services. The definition of the API Plugin includes the details of the external API service, such as endpoints, parameters, authorization, etc., and the Agent can call according to the above definition; Parameters, in the above task-oriented dialogue method based on large model Agent, parameters refer to the key information required to complete the task flow, which is generally filled by analyzing the information mentioned by the user through the dialogue system. When processing dialogues, the Agent needs to obtain some key information to effectively execute specific task flows or tasks. The above key information is the parameter, which can be analyzed and filled into the corresponding task flow in the dialogue process. Parameters can include various types of information, such as date, time, location, product name, order number, etc. When the user mentions the above information in the dialogue, the dialogue system can be used to analyze and extract the above information, and pass it to the Agent as a parameter, so that the Agent can execute the corresponding task according to the above information;

[0003] Mermaid, a JavaScript-based diagramming tool, provides a simple Markdown-like syntax for generating flowcharts, sequence diagrams, Gantt charts, and other types of diagrams. Mermaid's setup allows users to create complex diagrams using textual descriptions without directly manipulating graphical editing tools. Mermaid's syntax is similar to a lightweight markup language (e.g., Markdown), where users can define the structure and elements of a diagram using simple text descriptions. By using specific syntax rules, users can define nodes, edges, styles, and layouts in a diagram. For example, users can use simple text descriptions to define flowchart nodes, arrows, and connections, allowing them to quickly create complex flowcharts. According to the embodiments of the present disclosure, a method for processing interactive information is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here. Considering the large amount of model parameters of the large model and the limited computing resources of the mobile terminal, the above-mentioned method for processing interactive information provided by the embodiments of the present disclosure can be applied to the application scenario as shown in FIG. 1, but is not limited thereto. In the application scenario as shown in FIG. 1, the large model is deployed in a server a, which can be a cloud. The server a can be connected to one or more client devices b through a local area network connection, a wide area network connection, an Internet connection, or other types of data networks. The client devices b here can include but are not limited to: smartphones, tablet computers, notebook computers, palm computers, personal computers, smart home devices, vehicle-mounted devices, etc., which collectively constitute the client opposite the server. An operation interface for obtaining input information to be detected can be deployed on a graphical user interface on the client device. The client device b can interact with the user through the graphical user interface to call the large model, thereby realizing the method for processing interactive information provided by the embodiments of the present disclosure. In the embodiments of the present disclosure, the client device and the server constitute a dialogue system, which can perform the following steps: if the user needs to reply to the corresponding question through the dialogue robot, the input information to be replied can be input on the operation interface on the client device. The client device can obtain the input information and send it to the server through the network.After the server receives the input information, the following steps can be performed: Step S102, obtaining input information to be replied; Step S104, obtaining configuration information associated with the input information, wherein the configuration information is used to describe an initial information generation rule required to be met in a process in which an information processing model generates reply information corresponding to the input information, the initial information generation rule is used to guide the information processing model to determine a target information generation rule required to be met in the process of generating the reply information; Step S106, generating prompt information of the information processing model based on the configuration information and pre-obtained parameter information associated with the input information, wherein the parameter information is used to define key content in the reply information; Step S108, generating the reply information including the key content by using the prompt information. The generated reply information corresponding to the input information can be output to the client device, and after the client device receives the reply information, the reply information can be displayed. It should be noted that in the case that the running resources of the client device can meet the deployment and running conditions of the large model, the embodiment of the present disclosure can be performed in the client device. In the above running environment, the present disclosure provides a processing method of interactive information as shown in FIG. 2. It should be noted that the processing method of interactive information of this embodiment can be performed by the mobile terminal of the embodiment shown in FIG. 1. FIG. 2 is a flow chart of a processing method of interactive information according to an embodiment of the present disclosure, as shown in FIG. 2, the method can include the following steps: Step S202, obtaining input information to be replied. In the technical solution provided in step S202 of the present disclosure, the input information can be query information (query), which can also be called inquiry information, for example, it can be a user expression, it can be content input by a user on an operation interface of a client device and requiring a dialogue robot to reply, for example, it can be a corresponding question of the user to the dialogue robot, such as inquiring weather, handling Electronic Toll Collection (ETC) matters, etc., wherein the user can be a C-end user. It should be noted that the above input information is only for illustration, and is not specifically limited here, as long as the application scenarios and specific content of the input information can meet the user's query requirements, which are within the protection scope of the embodiment of the present disclosure. In this embodiment, the input information to be replied by the user can be obtained. For example, if the user has a demand for the dialogue robot to answer and reply related questions, the user can perform a corresponding input operation on an operation interface of the client device for dialogue with the dialogue robot to input a user expression that meets the user's query and question demand.When the operation interface detects that there is input information that has been input, the input information can be transmitted to the corresponding server through the network, and the corresponding analysis is performed to obtain reply information that replies to the input information. In the technical solution provided in step S204 of the disclosure, the configuration information can be used to describe the initial information generation rule required to be satisfied by the information processing model in the process of generating the reply information corresponding to the input information, and can include flow information and plug-in information. The configuration information can be used to determine the generation rule of the process of generating the above-mentioned reply information, which is similar to a template, and can be a pre-configured information generation rule. The initial information generation rule can be used to guide the information processing model to determine the target information generation rule required to be satisfied in the process of generating the reply information, and can be used to obtain the target information generation rule through flow reasoning, that is, the information processing model can plan according to the task flow and guide the conversation. The target planning can be a newly planned flow based on the demand information of the application scene. The information processing model can be a large model agent. oThe information processing model can analyze the input information of the user and flexibly fill or change parameter values when necessary. Optionally, the process information can be configured through a process configuration process, also referred to as process configuration information, which can include name, description, and canvas information, also referred to as a canvas. The plug-in information can be API plug-in information, also referred to as API plug-in configuration information, which can include name, description, and input / output parameter description and other information. The above information can be provided to the information processing model, and the information processing model can autonomously plan according to the API plug-in information to obtain target information generation rules. It should be noted that the information contained in the above process information and plug-in information is only for illustration and is not specifically limited herein. In this embodiment, after obtaining the input information to be replied, the configuration information associated with the input information can be obtained. Optionally, the operation personnel can configure the process configuration information and the API plug-in configuration information, for example, the operation personnel can configure the above two types of configuration information through a corresponding management page, wherein the operation personnel can also be referred to as a B-end operation personnel. Optionally, the management page for configuring the process information provides a direct and intuitive way for the B-end operation personnel to describe the task process. The nodes and connections in the canvas information describe the task process, and the dialog system can compile the above nodes and connections into a Mermaid syntax structure for processing by the large model agent. The above setting helps to simplify the process of configuring the process and enables the dialog system to better understand and process the task process, wherein the dialog system can be a task-oriented dialog system. Due to the complexity of the process configuration of the canvas in the related art, which is difficult to maintain, mainly due to the complexity of the dialog, the randomness of the dialog, and the flexible jump requirements, as well as the limitations of NLU in the related art. In view of the above problems, it is necessary to seek a more intelligent, flexible, and easy-to-maintain dialog system setting and configuration method. Therefore, in the embodiments of the present disclosure, the canvas does not need to define the dialog logic manually, but can define the core task process. Relying on the reasoning ability and instruction compliance of the large model, the above setting can greatly simplify the configuration of the nodes and edges on the canvas, making the configuration more concise and intuitive, thereby achieving the technical effect of simplifying the process configuration of the canvas. Optionally, the API plug-in configuration is a key function, through which the B-end operation personnel can configure the API plug-in information on the management page, including but not limited to name, description, and input / output parameter description. The above information will be provided to the large model, which can autonomously plan according to the API plug-in name and description and call the API plug-in when necessary to obtain the result.Subsequently, the large model can continue planning based on the result and generate the final reply information. Optionally, after the server obtains the input information from the client device, the input information can be subjected to process / API plug-in coarse recall, so as to obtain process information and plug-in information in the configuration information associated with the input information, wherein the process / API plug-in coarse recall can also be referred to as API+process recall. In step S206, prompt information of the information processing model is generated based on the configuration information and the pre-obtained parameter information associated with the input information. In the technical solution provided in the above step S206 of the present disclosure, the parameter information can be used to define the key content in the reply information, that is, the key information definition in the dialogue question and answer link of the user and the dialogue robot, or the key and core parameters required by the dialogue robot, which can also be referred to as parameter, parameter configuration information and parameter value, and can include name and description information. The parameter information can be used to determine the key semantic content in the final reply information, and the parameter value of the parameter information can be filled or changed flexibly according to different input information of the user. That is, the parameter information can change according to the input information input by the user. The parameter information can be generated by the dialogue system analyzing the information input by the user, that is, the input information, analyzing and understanding the intention of the user, to generate the key content in the final reply information. For example, if the parameter information analyzes from the input information that the user's intention is to develop a photo, the key content in the final reply information can be an inquiry about the number of developed photos. It should be noted that the key content in the reply information generated by the above parameter information and the corresponding input information are only for illustration and are not specifically limited herein. The prompt information can be a Prompt. The key content can also be referred to as key information. It should be noted that the above parameter information is only for illustration and is not specifically limited herein. In this embodiment, after obtaining the configuration information associated with the input information, the prompt information of the information processing model can be generated based on the configuration information and the parameter information corresponding to the reply information. Optionally, parameter configuration is an important part of the dialogue system. The parameter is the key information definition in the dialogue question and answer link, which usually contains name and description information. The large model can analyze the user's expression and flexibly fill or change the parameter value when necessary. For example, for the parameters of ETC transaction, vehicle information, user identity verification information, traffic rules, etc. It should be noted that the above parameters are only for illustration and are not specifically limited herein and can be set according to the dialogue needs of actual users. Optionally, the parameter information configured by the B-end operator is called. Through the above configuration information and parameter information, the Prompt can be assembled.and sending the Prompt to the large model Agent. oIn the embodiments of the present disclosure, the setting of the Prompt structure covers task description, process information, API plug-in information, and parameter information, and the like, which are combined together to call the large model. By combining the above information, the Prompt structure provides rich information required by the large model, helps the large model better understand the user's intention, perform the corresponding task, and generate accurate replies, so as to achieve the goal of intelligent dialogue, thereby achieving the technical effect of improving the accuracy of dialogue between the dialogue robot and the user. In the technical solution provided by the step S208 of the present disclosure, the reply information can be reply information (Answer) that replies to the input information input by the user. The reply information can be a human-like reply, that is, a reply expressed in a human language expression manner, which can also be referred to as a final reply. The reply information can include key content. The key content can be an answer source, that is, a key point used to answer the user expression analyzed from the user expression. For example, if the user expression input by the user is "Hello, has the Rizhao toll station been closed? I want to check the road condition information of the G1511 Rizhao section", the key content contained in the reply information can be "Answer source: 1, road condition, service area, toll station query" by analyzing the above user expression. It should be noted that the above key content is only an example for illustration, and is not specifically limited here, as long as the content can reflect the answer source of the current reply information, which is within the protection scope of the embodiments of the present disclosure. In this embodiment, after generating the prompt information of the information processing model based on the configuration information and the pre-acquired parameter information associated with the input information, the prompt information can be used to guide the information processing model to generate reply information corresponding to the input information according to the initial information generation rule and the target information generation rule. Optionally, after the large model Agent receives the Prompt, it can perform autonomous planning based on the above configuration information and parameter information, realize flexible jumping within the process, and dynamic calling capability of API, which can make the large model Agent better understand the user's intention and more accurately schedule between the process and API plug-in. By using the above large model Agent capability, a human-like reply can be generated more smoothly and naturally, that is, the reply information consistent with the user's expression habit. Optionally, after the information processing model generates the reply information corresponding to the current input information, the reply information can be transmitted to the operation interface on the corresponding client device through the network for display.In related technologies, a dialogue system in a pipeline architecture can be used, which can include an NLU component, a DM component, and an NLG component. The user intent is analyzed and understood through the NLU component, the core parameters are extracted, the user behavior is parsed, the user dialogue state is maintained and updated through the DM component combined with the user intent, the behavior is decided based on the task flow, and the reply is generated according to the user's behavior in the round through the NLG. In actual task requirements, B-end operators need to describe the task flow by configuring the canvas, and the dialogue policy (Dialogue Policy, DP for short) component, the NLG component and other components need to refer to the task flow generation result. However, the above process is not only cumbersome for operators, but also poor in management flexibility of the entire dialogue system. Therefore, there is still a technical problem of low processing efficiency of interactive information. However, in the embodiments of the present disclosure, in view of the poor management flexibility of the dialogue system in the above related technologies, an end-to-end large model task-based dialogue solution can be set up, NLU, DM and NLG can be processed by a large model, and the API plugin capability of the large model can be integrated to greatly improve semantic understanding and dialogue flexibility. In terms of operation capability, the solution can provide a task flow-oriented minimalist canvas interaction scheme for platform users, which facilitates them to quickly configure the task flow into a dialogue flow. Through the above method, by integrating the large model to process NLU+DM+NLG tasks, integrating the API plugin capability of the large model, and providing a task flow-oriented minimalist canvas interaction scheme, the performance, flexibility and user friendliness of the dialogue system can be greatly improved, which brings new opportunities for the development of task-based dialogue systems, thereby achieving the technical effect of improving the processing efficiency of interactive information. Through the above steps S202 to S208 of the present disclosure, if it is necessary to reply to the input information to be replied, the input information can be obtained, and the input information can be input into the information processing model to process the input information. In the processing process, the initial information generation rule required to be met by the input information, that is, the configuration information of the information processing model, can be determined. Based on the configuration information and the corresponding parameter information obtained in advance, the prompt information for guiding the information processing model to process the input information and generate the reply information can be generated. Through the prompt information, the target information generation rule required to be met in the process of generating the reply information by the information processing model is determined. Therefore, the reply information containing the key content can be generated according to the target information generation rule.In view of the problem that too much participation of an operator in configuration of a dialogue system leads to low efficiency of processing of interactive information, an embodiment of the present disclosure can define configuration information related to a task flow of processing of interactive information in the dialogue system, which is easy to master and maintain, and can effectively avoid giving a large amount of interference information to an information processing model. Based on this, the embodiment of the present disclosure relies on inference capability and instruction compliance of the information processing model, uses prompt information generated by the configuration information, so that the information processing model can determine a target information generation rule in the process of generating reply information, that is, autonomously plan a flow, without defining a corresponding dialogue logic artificially, thereby achieving the purpose that the information processing model can smoothly and naturally generate reply information of the reply input information, and further achieving the technical effect of improving the processing efficiency of interactive information, and solving the technical problem of low processing efficiency of interactive information. The above method of the embodiment is further introduced as follows. As an optional implementation manner, the initial information generation rule includes an initial execution flow and / or an initial calling interface; and the target information generation rule includes a target execution flow and / or a target calling interface. In the embodiment, the configuration information can be used to describe an initial execution flow required by the information processing model in the process of generating the reply information. The configuration information can also be used to describe an initial calling interface required by the information processing model in the process of generating the reply information. The initial execution flow and the initial calling interface described above can be used to guide the information processing model to determine a target execution flow required in the process of generating the reply information. The initial execution flow and the initial calling interface described above can be used to guide the information processing model to determine a target calling interface required in the process of generating the reply information. The initial information generation rule can include the initial execution flow and the initial calling interface. The target information generation rule can include the target execution flow and the target calling interface. Optionally, the initial execution flow can be flow configuration information, which can also be referred to as flow information. For example, it can be a node and an edge (a connection line) of a core task flow in a canvas, that is, the task flow can be described by using the node and the connection line in the canvas, where the task flow can also be referred to as a core task flow. The initial calling interface can be API plug-in information. The target calling interface can be an external resource, for example, an external API plug-in. In the embodiment, an operator at a B end can configure flow information through a management page, that is, the operator can input, edit and manage related information of a task flow in a related dialogue system. The above information can include a name, a description, a node, a connection line and the like of the flow. In addition to the name and the description, the more important thing in the flow information is canvas information.The canvas information can include a graphical representation of the task flow, such as the layout and connection relationship of nodes and lines. The operator can describe the task flow in the canvas by combing the task flow using nodes and lines, that is, the operator needs to comb the task flow and describe the task flow using nodes and lines on the canvas. The above nodes and lines represent different task steps or associations between processes. Alternatively, the nodes and lines in the canvas flow can be compiled into a Mermaid syntax structure by the dialogue system, which can indicate that the dialogue system converts the nodes and lines in the canvas into a specific Mermaid syntax structure. Mermaid is a text description language for flowcharts and sequence diagrams, which can convert graphical flow information into programmable text structures, that is, Mermaid syntax structure is a concise syntax for describing flowcharts, which expresses the nodes, edges and layout in the flowchart through text description. In the above setting, the B-end operator can describe the task flow through the canvas, and the dialogue system can generate the corresponding Mermaid syntax structure according to the description. The above setting helps to make the description of the task flow more intuitive and flexible, and facilitates the processing of the task flow by the dialogue system. Through the above configuration mode, the B-end operator can more intuitively comb the task flow and configure the process in the management page. The nodes and lines in the canvas information can be compiled into a Mermaid syntax structure by the dialogue system, which makes the task flow configuration more intuitive and enhances user friendliness. At the same time, the dialogue system can better understand and process the above task flow, providing a clearer input for the processing of the large model agent. Alternatively, the compiled Mermaid syntax structure is saved for processing by the large model agent. In related technologies, the complexity of the flow configuration of the canvas is difficult to maintain mainly due to the complexity of the dialogue, the randomness of the dialogue and the demand for flexible jump, and the limitations of NLU. In view of the above problems, it is necessary to seek a more intelligent, flexible and easy-to-maintain dialogue system setting and configuration mode. Due to the complexity of the application dialogue itself, it may be necessary to continuously add new nodes and lines in the canvas to support. As the task demand and dialogue scene continue to expand, the canvas configuration mode may not be able to cope with complexity well, resulting in canvas configuration being too bloated and difficult to manage. The dialogue itself has randomness and needs to support flexible jump. DM may not be able to completely cover the above flexibility, resulting in the need to continuously add lines to solve the jump problem in the dialogue flow, making the canvas configuration complex and difficult to maintain. NLU may have limitations, resulting in the need to write special processing logic in the node for secondary processing.The above situation can lead to the logical complexity of the node, increase the complexity and difficulty of the configuration. In summary, the related art still has the technical problem of high complexity of the flow configuration of the canvas. However, in the canvas design scheme provided by the embodiments of the present disclosure, all the dialogue logics do not need to be defined artificially, but only the core task flow needs to be defined. Relying on the reasoning ability of the large model and the instruction compliance, the above design can greatly simplify the configuration of the nodes and edges on the canvas, so that the configuration is more concise and intuitive. Specifically, in the above design scheme, the nodes and edges on the canvas are very concise, and the speech and similar questions on the edge can be directly edited for the large model to plan the flow, so that the configuration of the task flow is more intuitive and flexible. In addition to supporting direct editing of the reply on the node, external API plug-ins and parameter collection can also be called. The above design makes the configuration of the node more flexible and can meet the diversified task requirements. Under the advanced model, the node also supports directly describing the instruction in a natural language manner, and the large model can follow the instruction to guide the progress of the dialogue. The above design makes the configuration of the dialogue logic more intelligent and flexible, which conforms to the intelligent reasoning and instruction compliance ability of the large model, and the advanced model can be a model with more powerful functions and capable of completing more complex tasks. In summary, the canvas design scheme in the embodiments of the present disclosure greatly simplifies the configuration process of the dialogue system by means of the reasoning ability of the large model, through concise node and edge configuration, support for natural language description of instructions and other ways, thereby improving the intuitiveness and flexibility of the configuration, and bringing new opportunities for the intelligentization and personalization of the dialogue system, thereby realizing the technical effect of reducing the complexity of the flow configuration of the canvas. Optionally, the above canvas design scheme makes the user only need to describe the task flow in the canvas and call external API plug-ins when necessary, so as to complete the configuration of the flow canvas. The above setting effectively reduces the learning cost of the user and improves the ease of use of the canvas flow. The user can describe the task flow in an intuitive way without needing to understand complex technical details or grammar rules, thereby reducing the learning cost of the configuration. Through simple operations, the user can define the nodes, edges and calling of external API in the task flow, and complete the configuration of the entire dialogue system. At the same time, by calling the external API plug-in, the user can conveniently integrate external services and data sources, thereby enriching the functions of the dialogue system. The above design enables the user to more flexibly customize and configure the dialogue system to meet the requirements of different task scenarios. In summary, the above canvas design scheme simplifies the configuration process and integrates external API plug-ins, effectively reduces the learning cost of the user, improves the ease of use of the canvas flow, and provides a more intuitive, flexible and intelligent configuration method for the user.In this embodiment, the B-side operator configures the API plugin information on the management page, and the large model plans autonomously based on the API plugin information, and calls the API plugin to obtain the result when necessary, and generates the final reply according to the result. Optionally, the B-side operator can configure the API plugin information on the management page, including the plugin name, description, and input / output parameter specification. The configuration of the above information can enable the large model to understand the available API plugins and the functions and services that the API plugins can provide during the conversation. When the user converses with the large model, the large model can obtain the API plugin information through the Prompt based on the user's input and the context of the current conversation. The large model can autonomously plan based on the name and description of the API plugin to determine whether to call a certain API plugin to obtain the result. If the large model believes that the API plugin needs to be called, it can send a request to the corresponding plugin to obtain the result. After the large model obtains the result of the API plugin, it can continue to plan based on the result to generate the final reply and return it to the user. The above design enables the large model to call external API plugins when necessary, thereby expanding its functions and services and enabling it to more intelligently handle user requests and problems. In general, the above design fully utilizes the functions of external API plugins, enabling the large model to more flexibly and intelligently plan conversations and providing more rich services and functions. In the embodiments of the present disclosure, through the configuration of the API plugin information process, the operator can flexibly configure the API plugin information, providing the large model with more external data and service support. The large model can autonomously plan and decide based on the API plugin information, call the API plugin to obtain the required result when necessary, and then use the result for the final conversation reply generation. Through API plugin configuration, the conversation system can more intelligently utilize external resources to provide more rich and personalized conversation services. The above design can effectively improve the flexibility and intelligence of the conversation system, providing users with a more expected conversation experience. As an optional implementation, the similarity between the input information and the configuration information is greater than the similarity threshold. In this embodiment, in the configuration information associated with the input information, the configuration information with a similarity greater than the similarity threshold can be obtained, where the similarity threshold can be a pre-set value or a value set according to actual conditions, for example, it can be pre-set to 98%.The similarity threshold value can also be the number of configuration information with high similarity, for example, five configuration information with high similarity in the configuration information can be pre-set. It should be noted that the setting method, value and number of configuration information of the similarity threshold value are only for illustration, and are not limited herein. As long as the process and method of finding similar configuration information of the input information are within the protection scope of the embodiments of the present disclosure. In the embodiments of the present disclosure, in order to ensure the prediction accuracy of the large model and reduce the prediction cost, the configuration information with high similarity to the input information to be replied can be recalled, that is, the corresponding number of process information and API plug-in information with similar similarity to the user expression can be calculated from a large number of configuration information through the judgment of the similarity size. The above process means that the process or API plug-in similar to the user expression can be found from the dialogue system, so as to quickly find the related configuration information from a large number of configuration information, thereby realizing the technical effect of improving the efficiency of screening configuration information. As an optional implementation, obtaining the configuration information associated with the input information includes: performing vectorization processing on the input information to obtain a first vector; querying, from a vector database, a second vector with a similarity to the first vector greater than a similarity threshold value, wherein the vector database is configured to store a plurality of vectors, and different vectors are configured to represent results of vectorization processing on different configuration information; and determining configuration information corresponding to the second vector. In the embodiment, in the process of obtaining the configuration information with a similarity to the input information greater than the similarity threshold value, the input information can be vectorized to obtain the first vector after vectorization processing. The second vector with a similarity to the first vector greater than the similarity threshold value can be queried from the vector database. The configuration information corresponding to the second vector can be determined as the configuration information corresponding to the current input information, wherein the first vector can be a vector obtained by vectorizing the input information. A large number of vectors can be stored in the vector database, and different vectors can be used to represent the results of vectorization processing on different configuration information. The second vector can be used to represent the corresponding result of vectorization processing on the configuration information in the vector database, that is, the vector obtained by vectorizing the process information and API plug-in information pre-stored in the vector database. Vectorization processing can be referred to as vectorization, which is a process of converting text information into a vector form. The semantic and feature of the text information can be represented by using word embedding technology.In the embodiments of the present disclosure, in a dialogue system, a vector database can be used for coarse recall of process / API plug-ins, that is, process information and API plug-in information with high similarity to the current input information can be filtered out from the vector database to improve the prediction accuracy of the large model and reduce the prediction cost. Optionally, after the B-end operation personnel management page configures the process information and API plug-in information, the above information can be vectorized and saved in the vector database. The purpose of the above method is to enable the computer to better understand and process the text information corresponding to the process information and API plug-in information. Optionally, when the C-end user has a dialogue, the user's expression is also subjected to vectorization processing, and then the vector database is called. The vector database can calculate process information or API plug-in information with a similarity higher than a similarity threshold to the user's expression by using a vector similarity algorithm, for example, 5 process information or API plug-in information close to the user's expression can be determined by using the vector similarity algorithm. That is, the dialogue system can find process or API plug-in information similar to the user's expression, so as to quickly find the possible relevant content from a large amount of information. It should be noted that the above vectorization processing and the process of filtering out process information and API plug-in information close to the user's expression are only for illustration and are not limited in this regard. Optionally, using the vector database for coarse recall of process / API plug-ins can discard other process / API plug-ins irrelevant to the user's expression to a great extent, avoiding a large amount of irrelevant information from being given to the large model. The above method reduces the model prediction cost, because the large model does not need to process a large amount of irrelevant information, and also improves the model prediction accuracy, avoiding the model from mistakenly recalling irrelevant processes or API plug-ins. In general, the coarse recall of process / API plug-ins by the vector database can effectively reduce the interference of irrelevant information, improve the efficiency and prediction accuracy of the model, and thus provide more accurate and fast dialogue services for users. As an optional implementation, in the vector database, the second vector with a similarity greater than a similarity threshold to the first vector includes: determining a scenario to which the first vector belongs; and querying, in a vector set corresponding to the scenario in the vector database, the second vector with a similarity greater than the similarity threshold to the first vector, wherein the vector set includes at least one vector corresponding to a process belonging to the scenario and / or at least one vector corresponding to an interface belonging to the scenario. In this embodiment, in the process of querying the second vector with a similarity greater than the similarity threshold to the first vector in the vector database, the scenario to which the second vector belongs can be determined.In the vector set corresponding to the scene in the vector database, the similarity between the first vector and the second vector greater than the similarity threshold is queried, wherein the vector set can include at least one vector corresponding to the execution flow of the scene, and the vector set can also include at least one vector corresponding to the calling interface of the scene. The vector database can include vector sets corresponding to different scenes respectively. The scene can be a task that can be coped with by the large model, such as a high-speed information query scene or a photo washing scene. It should be noted that the above-mentioned scene is only for illustration, and is not limited herein, as long as the scene can be coped with and answered by the user expression by using the large model, which is within the protection scope of the embodiments of the present disclosure. Optionally, each configuration information can be identified as belonging to a specific scene. The above-mentioned scene can be a predefined label or a unique task identifier. In the vector database, the vector of each configuration information should be associated with the task identifier of the corresponding scene. Optionally, after the user expression is vectorized, the dialog system can calculate the similarity between the first vector and the vectors in the vector set corresponding to each scene in the vector database. Common similarity calculation methods include cosine similarity, Euclidean distance, etc. The above-mentioned calculation methods can measure the similarity between two vectors. It should be noted that the above-mentioned similarity calculation method is only for illustration, and is not limited herein. Optionally, according to the result of the similarity calculation, the dialog system can select several second vectors with high similarity to the first vector of the user expression. The above-mentioned second vector usually represents a scene and corresponding configuration information that are more relevant to the user expression. As an optional implementation, determining the configuration information corresponding to the second vector includes: in response to the second vector including a vector corresponding to the execution flow of the scene, determining the execution flow of the scene as an initial execution flow required to be executed in the generation process of the reply information to obtain the configuration information; and in response to the second vector including a vector corresponding to the calling interface of the scene, determining the calling interface of the scene as an initial calling interface required to be called in the generation process of the reply information to obtain the configuration information. In this embodiment, in the process of determining the configuration information corresponding to the second vector, if the second vector includes a vector corresponding to the execution flow of the scene, the execution flow of the scene can be determined as an initial execution flow required to be executed in the generation process of the reply information to obtain the configuration information. If the second vector includes a vector corresponding to the calling interface of the scene, the calling interface of the scene can be determined as an initial calling interface required to be called in the generation process of the reply information to obtain the configuration information.Optionally, after determining the second vector with high similarity to the first vector, the dialogue system can determine the task scene to which the above vector belongs, query the configuration information corresponding to the second vector and the belonging scene from the database. Optionally, in the process of determining the configuration information corresponding to the second vector from the database, it can be determined whether there is corresponding process information under the scene, if yes, the process information can be determined as the initial execution process of the large model, otherwise, it can be determined that there is no process information under the scene. It can also be determined whether there is a corresponding API plug-in under the scene, if yes, the API plug-in can be determined as the initial calling interface of the large model, otherwise, it can be determined that there is no API plug-in under the scene. The process information and / or API plug-in possessed by the above scene can be determined as the configuration information. Optionally, the dialogue system can calculate the similarity between the first vector of the user expression and the second vector in the vector database, and find the scene with high relevance and the corresponding process and / or API plug-in, so as to provide the service and information required by the user. In the embodiment of the present disclosure, the dialogue system can convert the above configuration information into a vector form by vectorizing the process and API plug-in information, and save it in the vector database. When the C-end user makes an expression, the user expression can also be processed by vectorization, and the vector database is called. The vector database finds 5 process or API plug-in information close to the user expression by calculating the vector similarity. The above setting enables the dialogue system to quickly find information related to the user expression, improves the efficiency and accuracy of the dialogue system. Through the above manner, the dialogue system can discard other processes or API plug-ins irrelevant to the user expression to a great extent, avoiding bringing a large amount of interference information to the model. The above manner helps to reduce the prediction of the model and improves the prediction accuracy of the model. In general, by vectorizing the process and API plug-in information and using the vector database for rough recall, the dialogue system can more quickly and accurately find information related to the user expression, achieving the technical effects of improving the efficiency and accuracy of the dialogue system. As an optional embodiment, step S206, based on the configuration information and the pre-acquired parameter information associated with the input information, generates prompt information of the information processing model, including: obtaining parameter information from a server; combining the configuration information and the parameter information to obtain the prompt information of the information processing model. In this embodiment, in the process of generating the prompt information of the information processing model based on the configuration information and the pre-acquired parameter information, the parameter information can be obtained from the server, and the configuration information and the parameter information can be combined to obtain the prompt information of the information processing model, wherein the server can be a B-end.In this embodiment, the importance of parameter configuration in the dialogue question and answer link is very high, therefore, how to define the parameter and how the large model processes the parameter value of the parameter is very important. Optionally, the parameter plays a key role in the dialogue question and answer link, and can be used to define the key information required in the dialogue. The configuration of the parameter includes the name and description information of the parameter, which can help the large model understand the content required to be collected and processed in the dialogue. When the user conducts the dialogue, the large model can analyze the user's expression, identify the information provided by the user, and flexibly fill or change the parameter value according to the context of the dialogue. The above process means that the large model has the ability to intelligently identify and process parameters, and can dynamically update the value of the parameter according to the user's input, so as to better understand the user's intention and generate a corresponding reply. Through the parameter configuration, the large model can better understand and process the key information in the dialogue, making the dialogue more intelligent and flexible. The flexible configuration and processing of the parameter enable the large model to better adapt to different dialogue scenarios and user needs, thereby providing a more personalized and customized dialogue experience. In the present disclosure, through the above parameter configuration, the dialogue system can better understand and process the key information in the dialogue. The large model can identify and process the above parameters during the dialogue process, so as to perform corresponding operations according to the user's requirements. The above setting can make the dialogue system more intelligent and personalized, and improve the understanding and response ability of the dialogue system to user needs. Therefore, the parameter configuration is an indispensable part of the dialogue system, which can help the dialogue system better understand the user's intention, thereby realizing more intelligent and accurate dialogue interaction. In this embodiment, the prompt information can be combined through the above configuration information and parameter information, that is, the prompt can be assembled through the above flow with high similarity, API plug-in, and parameter information obtained from the B side. As an optional implementation, the method further includes: obtaining at least one of the following auxiliary information: task description information in a scene to which the input information belongs, environment information in the scene to which the input information belongs, and context information of the input information; and combining the configuration information and the parameter information to obtain prompt information of the information processing model, including: combining the configuration information, the parameter information, and the auxiliary information to obtain the prompt information. In this embodiment, part of the auxiliary information can be obtained to combine the configuration information, the parameter information, and the auxiliary information to determine the prompt information, wherein the auxiliary information can include task description information in a scene to which the input information belongs, environment information in the scene to which the input information belongs, and context information of the input information.Optionally, the task description information can also be referred to as a task description, which includes describing the role that needs to be played by the large model agent, the task to be performed, defining the format of input and output, and examples. The above part of information provides the specific task that needs to be performed by the large model and the expected input and output format, helping the large model to clearly understand the task requirements and expected results. Optionally, the process information is obtained by vector database calculation, and the process information is mainly the core task process described by the operation personnel through the canvas, which is expressed by using the Mermaid syntax structure. The large model can refer to this task process for planning and guiding the progress of the dialogue. Optionally, the API plug-in information is obtained by vector database calculation, and the API plug-in information inside is the external API plug-in information configured by the operation personnel. The large model can select to call the API when necessary, and then plan again according to the API result. Optionally, the parameter information is the key information in the dialogue process configured by the operation personnel. The large model can extract the parameter information in combination with the context, and understand based on the parameter information when necessary to complete the dialogue task. Optionally, the environment information can include the current time and the like, which can be used to assist the identification and planning of the large model. Optionally, the context information can be the multi-round conversation information of the C-end user, which can also be referred to as multi-round dialogue information, which can include multi-round user expressions and dialogue system replies, and can be used to express the context state. The above part of information is used to help the large model understand the context of the dialogue and maintain the coherence of the dialogue. Optionally, the setting of the Prompt structure covers the task description, the process information, the API plug-in information, the parameter information, the environment information and the context information, and the above information is combined to call the large model. By combining the above part of information, the Prompt structure provides rich information required by the large model, helping the large model to better understand the user's intention, perform the corresponding task and generate accurate replies, so as to achieve the goal of intelligent dialogue. In the embodiments of the present disclosure, the Prompt is assembled by combining the above-mentioned various auxiliary information, configuration information and parameter information, so as to provide the large model with the required rich information in multiple aspects, so as to make the large model more comprehensively understand the user's intention, thereby achieving the technical effect of improving the accuracy of the large model in analyzing the user's intention. As an optional implementation manner, in step S208, the reply information including the key content is generated by using the prompt information, including: the control information processing model determines the intention data of the input information, wherein the intention data is used to describe the purpose expressed by the input information; according to the intention data, the prompt information is used to guide the above-mentioned information processing model to generate the reply information according to the initial information generation rule and the target information generation rule.In the embodiment, in the process of generating the reply information by using the prompt information, the information processing model can be controlled to determine the intention data of the input information. According to the intention data, the information processing model can be guided to generate the reply information according to the initial information generation rule and the target information generation rule by using the prompt information, wherein the intention data can be used to describe the purpose expressed by the input information, that is, can be used to represent the purpose of the user's dialogue with the dialogue robot, and can also be called user intention. Optionally, after the Prompt is assembled, the Prompt can be transmitted to the large model Agent. After receiving the Prompt, the large model Agent can autonomously plan based on the process and API information, combined with the context state. The above process means that the large model can flexibly jump within the process, call the corresponding API plug-in, and dynamically process the user's request, so as to realize more intelligent and flexible dialogue interaction. Optionally, the semantic understanding ability of the large model is enhanced, which can understand the user's intention more deeply and schedule more accurately between the process and the API plug-in. The above process means that the large model can better understand the user's demand, so as to provide more personalized and accurate services. Optionally, the large model can understand the complete task process, so as to realize flexible jumping within the process and between processes. The above process means that even if the dialogue process itself is unpredictable, the large model can flexibly plan and control the process according to the user's demand and context state. Optionally, in addition, the large model also supports the user's multi-intention expression, and can intelligently clarify the intention or simultaneously execute multiple processes. The above process means that even if the user's expression is complex and diverse, the large model can process and understand the complex intention, and make corresponding responses and behaviors. Optionally, the large model can generate a human-like reply smoothly and naturally, avoiding the problem of fixed and rigid reply in related dialogue systems, thereby improving the naturalness and authenticity of dialogue experience. In the embodiment of the present disclosure, by the above manner, the large model Agent has more intelligent, flexible and natural characteristics in the dialogue system, can better adapt to the user's demand, provide more personalized and efficient dialogue interaction, thereby realizing the technical effect of improving the naturalness of the dialogue process. As an optional implementation manner, according to the intention data, the prompt information is used to generate the reply information, comprising: controlling the information processing model to determine the parameter value of the parameter information in the prompt information according to the intention data; according to the parameter value of the parameter information, the prompt information is used to generate the reply information.In the embodiment, in the process of guiding the information processing model according to the prompt information, generating the reply information according to the initial information generation rule and the target information generation rule, the information processing model can be controlled to determine the parameter values of the parameter information in the prompt information according to the intention data. The information processing model can be guided according to the parameter values of the parameter information, using the prompt information to generate corresponding reply information according to the initial information generation rule and the target information generation rule, wherein the parameter values can be parameter values of the parameter information corresponding to different scenarios, such as the parameter values of the sunshine toll station, the sunshine road section, and the G1511 expressway in the high-speed information query system scenario. It should be noted that the above parameter values are only for illustration, and are not specifically limited herein, as long as the parameter values can be determined according to different scenarios, which are within the protection scope of the embodiments of the present disclosure. Optionally, the large model can analyze the input information and the context of the user, recognize and understand the intention of the user. According to the analyzed intention of the user, the large model can identify the parameter information that needs to be changed and make corresponding parameter value changes. For example, the previously defined parameters can be updated or reset to ensure that the parameter values are consistent with the user's intention. Optionally, according to the changed parameter values, the large model can use Prompt guidance to generate a reply. Through Prompt, some suggestive information can be provided to the large model to guide the model to generate a specific reply. The above Prompt can include dialogue context, user intention, parameter information, etc., to help the model more accurately understand user requirements and generate corresponding replies. Optionally, based on the changed parameter values and using Prompt guidance, the large model can plan a dialogue and generate a reply. The large model can combine parameter information, user intention, and context state to generate natural and smooth replies to user requirements. By using Prompt guidance, the large model can be guided to more accurately interact in a dialogue, thereby providing replies that are more in line with user expectations. In the embodiments of the present disclosure, the large model can change the parameter values of the parameter information according to the analyzed intention of the user, and use Prompt guidance to generate a reply to the user's intention, thereby realizing more intelligent and flexible dialogue interaction. As an optional implementation, the information processing model is controlled to determine the parameter values of the parameter information in the prompt information according to the intention data, including: in response to the initial information generation rule including an initial execution process, the information processing model is controlled to determine the parameter values of the parameter information matched with the initial execution process according to the intention data.In the implementation example, in the process of determining the parameter value of the parameter information in the prompt information according to the intention data in the control information processing model, if the configuration information in the prompt information includes an initial execution flow, the control information processing model can be controlled to determine the parameter value of the parameter information matched with the initial execution flow according to the intention data, wherein the initial information generation rule can include the initial execution flow. In the implementation example of the present disclosure, the planning and memory capabilities of the large model can be utilized to significantly improve the flexibility and personification of the dialogue, and solve the rigidity of the DM and the non-personification of the NLG. The planning can be used to plan a more flexible and personalized dialogue flow according to the input and context of the user. The above process means that the large model can intelligently plan the dialogue according to the user's needs and intentions, thereby providing a more personalized and user-expected dialogue experience. The memory can remember the previous dialogue content and context information. The above process enables the large model to better understand the user's needs, the dialogue is more coherent, and consistency is maintained during the dialogue process, enhancing the naturalness of the dialogue. Optionally, the planning and memory of the large model enable the dialogue system to better adapt to different dialogue scenarios and user needs, providing more flexible and personalized dialogue services. The dialogue system can dynamically adjust the dialogue flow according to the user's input, thereby better meeting the user's needs. In related technologies, the DM can have rigidity, and the NLG can not generate a personified reply. Using the planning and memory capabilities of the large model, the dialogue can be made more flexible and natural, and more personified replies can be generated, thereby improving the realism of the dialogue and the user experience. In summary, the planning and memory capabilities of the large model can significantly improve the flexibility and personification of the dialogue, solving the rigidity of the related dialogue system and the non-personification problem of natural language generation, thereby providing users with a more intelligent, personalized, and natural dialogue experience. Optionally, in the corresponding scenario, because the user's expression can involve multiple intentions, the large model Agent can understand the multiple intentions described above and perform planning based on the flowchart to obtain the corresponding parameter value. For example, in the scenario of a high-speed information query system, it mainly includes query functions of road conditions, toll stations, and servers, and operating personnel can help users understand the high-speed situation in real time by configuring task flows. If the user's expression input in the first round of dialogue is "Hello, is the Rizhao toll station closed? I want to check the road condition information of the G1511 Rizhao section." In the above round of dialogue, the user's expression involves multiple intentions, and the large model Agent can understand and identify multiple intentions.After planning based on the flowchart, the large model agent collects three parameters, i.e., "Rizhao toll station", "G1511 highway", and "Rizhao section" in the above user expression, and sequentially executes the toll station information query and highway section query two APIs. Finally, a smooth and natural reply is generated by combining the results of the two APIs. The above process shows that the large model agent has strong semantic understanding and autonomous planning capabilities, can effectively handle complex multi-intent dialogues, and generate natural and smooth replies. The above capabilities enable the large model agent to better cope with the user's diversified and complex needs and provide a more intelligent and flexible dialogue experience. It should be noted that the above scenario and the specific process of the large model analyzing the user expression are only illustrative, and are not specifically limited herein. In the above embodiments of the present disclosure, the large model agent demonstrates its strong semantic understanding, autonomous planning, flexible jumping, and understanding and following complex instructions in the scenario of the highway information query system, thereby providing a smooth, natural, and efficient dialogue experience for the user. As an optional implementation, according to the parameter value of the parameter information, the reply information is generated by using the prompt information, including: controlling the information processing model to generate the reply information according to the initial information generation rule and the target information generation rule corresponding to the configuration information in the prompt information. In this embodiment, in the process of generating the reply information according to the parameter value of the parameter information and using the prompt information to guide the information processing model to generate the reply information according to the initial information generation rule and the target information generation rule, the information processing model can be controlled to use the configuration information corresponding to the parameter value in the prompt information to control the information processing model to generate the reply information. If the scene is a photo developing scene, the configuration information can be a recommended size process node. If the scene is a highway information query system scene, the configuration information can be a toll station information query API and a highway section query API. It should be noted that the above configuration information is only illustrative, and is not specifically limited herein, as long as the corresponding configuration information can be configured according to different scenes, which is within the protection scope of the embodiments of the present disclosure. Optionally, according to the user's input information and the context, the parameters that need to be filled or changed are identified, and the parameter values are matched with the pre-configured flow information and API plug-ins, i.e., the parameter values and the configuration information are compared or matched, and the corresponding matching result can be obtained to determine the process or API plug-in to be executed.Optionally, according to the matching result, corresponding prompt information is generated to prompt the large model of the process information and API plug-in that need to be executed, that is, through the process name, API plug-in name, and corresponding instructions or prompts included in the prompt information, the large model is helped to understand and execute the corresponding task process and operation. Optionally, according to the determined configuration information, the corresponding initial information generation rule and the target information generation rule are generated to guide the large model to generate the corresponding reply information. The above-mentioned rule is a pre-defined task logic and rule, which is used to guide the large model to generate a suitable reply under a specific condition. Optionally, the large model generates corresponding reply information according to the application of the prompt information and the rule. The above-mentioned reply information should meet the needs and expectations of the user, and can provide accurate, clear and task rule-compliant replies to meet the needs of the user. In the present disclosure, the large model can determine the process information and API plug-in corresponding to the parameter value according to the user's needs and context, and guide the large model to generate the expected reply information according to the determined configuration information and rule. The above-mentioned process implementation process can ensure that the dialogue system can provide accurate, flexible and task rule-compliant replies under different conditions. For example, in the scenario of developing photos, if the user's statement input in the first round of dialogue is "I want to develop photos". In the above-mentioned round of dialogue, the user explicitly indicates that he wants to develop photos. The large model Agent can understand the user's intention, and based on the task process configuration, it realizes that the number of photos is an important parameter, so it asks the user for information about the number of photos. The above-mentioned process shows that the large model Agent has the ability of task process understanding and parameter collection, and can flexibly collect necessary parameter information according to the task requirements and user intention. The above-mentioned ability enables the large model Agent to better understand the user's needs and provide personalized services for the user. In the present disclosure, the large model Agent demonstrates its ability of task process understanding, parameter collection, flexible jumping and intelligent reply generation in the above-mentioned dialogue case, thereby providing the user with a personalized and intelligent dialogue interaction experience. As an optional implementation manner, the control information processing model generates the reply information according to the initial information generation rule and the target information generation rule corresponding to the configuration information in the prompt information, including: in response to the initial information generation rule including initial execution, the control information processing model performs according to the initial execution process corresponding to the parameter value, to obtain a first execution result, wherein the initial information generation rule includes the initial execution process; the control information processing model determines the target information generation rule by using the first execution result; and the control information processing model performs according to the target information generation rule to obtain the reply information.In the embodiment, in the process of generating reply information according to the initial information generation rule and the target information generation rule corresponding to the determined configuration information in the control information processing model, if the configuration information can include an initial execution flow, the control information processing model can be used to execute the initial execution flow corresponding to the parameter value to obtain a first execution result. The control information processing model can use the first execution result to determine a target execution flow and a target calling interface. The reply information can be generated according to the target execution flow and the target calling interface. The initial information generation rule can include the initial execution flow. The target information generation rule can include the target execution flow and the target calling interface. The first execution result can be an execution result obtained by executing the initial execution flow, for example, a result of jumping to a consultation question reply node in a high-speed information query system scenario, or a result of jumping to a recommended size node in a photo printing scenario. It should be noted that the first execution result described above is only an example, and is not limited in this regard. Any first execution result that can be determined according to different scenarios is within the protection scope of the embodiments of the present disclosure. Optionally, after the first execution result is obtained, the corresponding target execution flow and target calling interface can be determined according to the first execution result, that is, scheduling can be performed between the flow and the API plug-in. For example, in the high-speed information query system scenario, if the user input in the second round of the question and answer is "when will it be opened", the user expression in the second round of the dialogue involves flow jumping, and the large model agent can flexibly jump to the consultation question reply node based on the user semantics and the flowchart, thereby avoiding complex connections within the canvas. The above process shows that the large model agent has the ability to flexibly jump within the node, and can flexibly plan the dialogue and control the flow according to the user demand and the context state. The above ability makes the large model agent more efficient and flexible when processing complex dialogue flows. For another example, in the photo printing scenario, if the user input in the second round of the question and answer is "10, what size can be recommended", after the user replies to the number of photos in the dialogue system in the above round of dialogue, the user expects the dialogue system to recommend a size. The large model agent flexibly jumps to the recommended size node based on the canvas configuration, and directly inquires whether the user needs the recommended size after recommending the size. The large model agent exhibits strong flow understanding and reply generation capabilities, and directly confirms with the user whether to use the recommended size in combination with the currently recommended size.The above process shows that the large model agent has the ability of flexible jumping and reply generation when processing complex task processes, and can intelligently plan the dialogue and generate replies according to the task rules and user needs. It should be noted that the above scenarios and specific user descriptions are only for illustration and are not limited herein. As an optional implementation, the control information processing model determines the target information generation rule based on the first execution result, including: the control information processing model determines the instruction description information in the target execution process based on the first execution result, wherein the instruction description information is used to describe the instructions allowed to be executed by the information processing model; and / or the control information processing model determines the target calling interface based on the first execution result. In this embodiment, in the process of determining the target execution process based on the first execution result by the control information processing model, the control information processing model can determine the instruction description information in the target execution process based on the first execution result. The control information processing model can also be used to determine the target calling interface of the corresponding target plug-in based on the first execution result, wherein the instruction description information can be used to describe the instructions allowed to be executed by the information processing model, which can be described in the form of natural language and can be used to express the operations required to be executed by the large model in steps. The target calling plug-in can be an external API plug-in. Optionally, in the canvas design scheme in the embodiments of the present disclosure, all dialogue logics do not need to be defined manually, but rely on the reasoning ability of the large model and the instructions followed. The nodes and edges on the canvas are very simple, and the edges can be directly edited with dialogues and similar questions, which are used for the large model to plan the process. In addition to supporting direct editing of replies on the nodes, external API plug-ins and parameter collection can also be called. Under the advanced model, the nodes also support describing instructions in the form of natural language, and the large model can follow the instructions to guide the dialogue. For example, in the ETC handling scene, the instruction description information can be the operations required to be executed by the large model in steps in this scene, such as first asking the user which type of ETC device is needed, including card-type ETC device and card-free ETC device. After the user selects the category, the corresponding device is introduced: 1. The card-type ETC requires the user to insert the ETC card into the vehicle-mounted card reader, and the vehicle-mounted card reader completes the fee deduction, and the corresponding price is 68 yuan; 2. The card-free ETC is directly integrated into the vehicle-mounted component (On-Board Unit, abbreviated as OBU) of the vehicle, and interacts with the highway toll station to achieve the purpose of automatic fee deduction. The card-free ETC is convenient and fast. It should be noted that the above instruction description information is only for illustration and is not limited herein.In the embodiments of the present disclosure, through the above setting, the user only needs to describe the task flow in the canvas and call external API plug-ins when necessary, and the configuration of the flow canvas can be completed. The above setting effectively reduces the learning cost of the user and improves the ease of use of the canvas flow. The user can describe the task flow in an intuitive way without understanding complex technical details or syntax rules, thereby reducing the learning cost of configuration. Through simple operation, the user can define the nodes, edges and external API calls in the task flow to complete the configuration of the entire dialogue system. At the same time, by calling external API plug-ins, the user can conveniently integrate external services and data sources, thereby enriching the functions of the dialogue system. The above setting enables the user to more flexibly customize and configure the dialogue system to meet the needs of different task scenarios. In general, by simplifying the configuration process and integrating external API plug-ins, the learning cost of the user is effectively reduced, the ease of use of the canvas flow is improved, and a more intuitive, flexible and intelligent configuration method is provided for the user. As an optional implementation, the control information processing model generates the reply information according to the determined initial information generation rule and target information generation rule corresponding to the configuration information, including: in response to the initial information generation rule including an initial call interface, the control information processing model calls the initial call interface for execution to obtain a second execution result; the control information processing model determines the target information generation rule by using the second execution result; and the control information processing model executes according to the target information generation rule to obtain the reply information. In this embodiment, in the process of generating the reply information by the control information processing model according to the determined initial information generation rule and target information generation rule corresponding to the configuration information, if the initial call interface is included in the configuration information, the control information processing model can call the initial call interface for execution to obtain the second execution result. The control information processing model can determine the target execution flow and the target call interface by using the second execution result. And the control information processing model executes according to the target execution flow and the target call interface to generate the reply information, wherein the initial information generation rule can include the initial call interface. The target information generation rule can include the target execution flow and the target call interface. The second execution result can be an execution result obtained according to the initial call interface, such as a result of jumping to the consultation question reply node in the high-speed information query system scenario, or a result of jumping to the recommended size node in the photo printing scenario. It should be noted that the above second execution result is only for illustration and is not limited in this regard, as long as the second execution result can be determined according to different scenarios, which is within the protection scope of the embodiments of the present disclosure.Optionally, after the large model receives the API plugin information, autonomous planning can be performed based on the names and descriptions of the plugins in the API plugin information. The above process means that the large model can identify the corresponding task flow or operation according to the API plugin information, and decide how to utilize the above plugins for processing in the dialogue. That is, the dialogue flow can be analyzed and decided to ensure that the dialogue system can reasonably call the API plugin to meet the user's needs. Optionally, when necessary, the large model can call the corresponding API plugin according to the result of autonomous planning to obtain the processing result. The information or service result provided by the API plugin can be obtained through the interaction between the dialogue system and the external API by sending a request and receiving a response. Optionally, after the large model obtains the result of the API plugin, planning can be continued based on the above result to adjust the dialogue flow and decision-making according to the result. The dialogue system can dynamically adjust the logic and process of the dialogue according to the return result of the API plugin to ensure that the final reply meeting the user's needs is generated. Optionally, the large model can generate the final reply based on the planning and the result of the API plugin. The above reply can include answering the user's question, providing the required information or service, etc. The large model can generate reply information meeting the user's expectations according to the result of the API plugin and the autonomous planning flow. In the embodiments of the present disclosure, through autonomous planning and result acquisition of the API plugin information, the large model can perform dialogue planning and flow control according to the plugin information, call the API plugin to obtain the result, and continue planning based on the result to finally generate a reply meeting the user's needs. The above process enables the dialogue system to more flexibly and intelligently handle various complex dialogue scenarios and user needs. For example, in the ETC handling scenario, if the user input in the third round of the question and answer is "What is the phone number?", the user continues to inquire about the toll station information in the third round of dialogue, and the large model Agent can understand multiple complex instructions in the node and reply according to the correct instructions. In addition, the rich instruction target configuration in the node can greatly reduce the number of nodes of the canvas. The above process shows that the large model Agent has the understanding and following ability of complex instructions, and also has flexibility in node configuration, which can better adapt to different task needs and simplify the configuration process. For another example, in the photo developing scenario, if the user input in the third round of the question and answer is "That's fine, 5 inches. I'll talk to you tomorrow, I'm a little busy today.", after the large model Agent collects the photo size, it can initiate API plugin calling and generate an order.However, when the user explicitly indicates that they are busy, the dialogue system does not continue to disturb the user, but ends the conversation in a friendly manner, and the order is determined to be valid for a long time. The above process demonstrates the intelligent and friendly conversation ending capability of the large model agent. As an optional example, the large model agent can automatically trigger the corresponding task flow, including API plugin calling and order generation, according to the user's needs. However, when the user indicates that they are busy, the large model agent can intelligently understand the user's attitude and decide not to disturb the user. The above process shows that the large model agent has intelligent dialogue processing capabilities and can adapt to the user's emotions and attitudes, thereby providing a more personalized dialogue experience. After the user indicates that they are busy, the large model agent can end the conversation in a friendly manner, and the order is determined to be valid for a long time. The above friendly conversation ending manner shows the importance of the large model agent in the customer service field scenario. The large model agent can help maintain a good user experience and avoid unnecessary disturbances, while showing the friendliness and respect of the dialogue system for the user. The large model agent can generate friendly and natural anthropomorphic replies in combination with the context. The above capabilities make the dialogue more natural and smooth, and improve the user's comfort and satisfaction during the dialogue. In the embodiments of the present disclosure, the large model agent also exhibits its intelligent dialogue processing capabilities, friendly conversation ending manner, and natural anthropomorphic reply capabilities. The above features enable the large model agent to provide more intelligent, personalized, and good experience dialogue services in the customer service field scenario. The embodiments of the present disclosure also provide another method for processing interactive information in an application scenario. FIG. 3 is a flowchart of another method for processing interactive information according to an embodiment of the present disclosure. As shown in FIG. 3, the method can include the following steps: Step S302, obtaining input information to be replied. In the technical solution provided in the above step S302 of the present disclosure, input information to be replied sent by a user can be obtained. Optionally, if the user has a need for a dialogue robot to answer and reply related questions, the user can perform a corresponding input operation on an operation interface on a client device that can dialogue with the dialogue robot, to input user expressions that meet the user's query and questioning needs. When the operation interface detects that there is input information that has been input, the input information can be transmitted to a corresponding server through a network. Step S304, determining a scene to which the input information belongs. In the technical solution provided in the above step S304 of the present disclosure, after obtaining the input information to be replied, the scene to which the input information belongs can be determined, where the scene can be a high-speed information query scene or a photo developing scene, etc. This is only an example for illustration and is not limited specifically.Optionally, the received input information to be replied can be analyzed and processed to determine the scenario to which the input information belongs. For example, the user expression input by the user can be classified and recognized in terms of intention by using natural language processing techniques. It should be noted that the above process and method of analyzing the scenario to which the user expression belongs are only for illustration and are not limited herein. In step S306, configuration information associated with the input information in the scenario is obtained, wherein the configuration information is used to describe initial information generation rules required to be met by the information processing model in the process of generating reply information corresponding to the input information, and the initial information generation rules are used to guide the information processing model to determine target information generation rules required to be met in the process of generating the reply information. In the technical solution provided by the above step S306 of the present disclosure, after determining the scenario to which the input information belongs, the configuration information associated with the input information in the scenario can be obtained. Optionally, after the server obtains the input information from the client device, the input information can be recalled by process / API plug-in, so as to obtain process information and plug-in information in the configuration information associated with the input information and adapted to the scenario. In step S308, prompt information of the information processing model is generated based on the configuration information and pre-obtained parameter information associated with the input information, wherein the parameter information is used to define key content in the reply information. In the technical solution provided by the above step S308 of the present disclosure, after obtaining the configuration information associated with the input information in the scenario, the prompt information of the information processing model can be generated based on the configuration information and the corresponding parameter information. Optionally, the parameter information configured by the B-end operator is called. By using the above configuration information and the parameter information, Prompt can be assembled. And the Prompt is sent to the large model Agent. oOptionally, the setting of the Prompt structure covers task description, process information, API plug-in information, and parameter information, and the above information is combined to call the large model. By combining the above information, the Prompt structure provides rich information required by the large model, helps the large model better understand the user's intention, perform the corresponding task, and generate accurate replies, so as to achieve the goal of intelligent dialogue, and thus realize the technical effect of improving the accuracy of dialogue between the dialogue robot and the user. In the technical solution provided by the step S310 of the present disclosure, after generating the prompt information of the information processing model based on the configuration information and the parameter information corresponding to the reply information, the prompt information can be used to guide the information processing model to generate the reply information corresponding to the input information according to the initial information generation rule and the target information generation rule. Optionally, after the large model Agent receives the Prompt, it can perform autonomous planning based on the above configuration information and parameter information, realize flexible jumping within the process, and dynamic calling capability of API, which can make the large model Agent better understand the user's intention and more accurately schedule between the process and API plug-in. By using the above large model Agent capability, a more fluent and natural human-like reply, i.e., reply information, can be generated. Optionally, after the information processing model generates the reply information corresponding to the current input information, the reply information can be transmitted to the operation interface on the corresponding client device through the network for display. Through the steps S302 to S310 of the present disclosure, the input information to be replied is obtained, the scene to which the input information belongs is determined, the configuration information associated with the input information in the scene is obtained, the prompt information of the information processing model is generated based on the configuration information and the pre-obtained parameter information associated with the input information, and the reply information meeting the scene and including the key content is generated by using the prompt information, thereby realizing the technical effect of improving the processing efficiency of interactive information and solving the technical problem of low processing efficiency of interactive information. The present disclosure also provides another method for processing interactive information from the side of the information processing device. FIG. 4 is a flowchart of another method for processing interactive information according to an embodiment of the present disclosure. As shown in FIG. 4, the method can include the following steps:In the technical solution provided in the foregoing step S402 of the present disclosure, the initial information generation rule required by the information processing model in the process of generating the reply information corresponding to the input information can be determined. Optionally, the task rule and logic of the dialogue model are defined, and the rule can include common dialogue scenarios, task processes, reply strategies, etc. The rule can guide the dialogue system to generate appropriate replies in different situations. According to the task requirements and user expectations, the dialogue strategy and rule are formulated. The dialogue strategy and rule can include standard answers to specific questions, rules for user question classification, processing rules for specific keywords or entities, etc. On the basis of the task rule, the initial dialogue process and decision tree are designed. The process can guide the dialogue system on how to generate replies in different situations, including how to handle specific questions and how to guide users, etc. By simulating and verifying the dialogue of some typical scenarios, the dialogue rule designed initially is verified and adjusted, which helps to ensure the accuracy and applicability of the rule. According to the task rule and dialogue strategy, some standard reply templates are formulated. The templates can be used as part of the initial information generation rule to guide the text structure and content of the large model when generating replies. By analyzing the feedback of the user and the replies generated by the dialogue system, the initial information generation rule is continuously adjusted and optimized, which helps to ensure that the rule can meet the user's needs and adapt to the changing dialogue scenarios. In the present embodiment, through the above steps, the initial information generation rule and dialogue strategy can be formulated to guide the large model to meet the task requirements and user expectations when generating the reply information corresponding to the user expression. The rule can guide the dialogue system to generate appropriate replies in different situations according to the task rule and dialogue strategy, thereby providing a dialogue experience that is more in line with the user's expectations. In step S404, configuration information is created based on the initial information generation rule, wherein the configuration information is used to generate prompt information of the information processing model, and the prompt information is used to generate the reply information. In the technical solution provided in the foregoing step S404 of the present disclosure, after the initial information generation rule required by the information processing model is determined, the configuration information can be created based on the initial information generation rule. Optionally, the B-end operator can configure the API plug-in information on the management page, including the name, description, and input / output parameter specification. The information will be provided to the large model, which can autonomously plan according to the API plug-in name and description and call the API plug-in to obtain the result when necessary. The B-end operator can configure the process information through the management page, which means that the operator can input, edit, and manage the related information of the task process in the relevant dialogue system.The initial information generation rule is used to guide the information processing model to determine the target information generation rule required to be met in the process of generating the reply information. Based on the initial information generation rule, configuration information is created, wherein the configuration information is used to generate prompt information of the information processing model, and the prompt information is used to generate the reply information, thereby achieving the technical effect of improving the processing efficiency of interactive information, and solving the technical problem of low processing efficiency of interactive information. The above method of the embodiment will be further introduced below. As an optional implementation manner, the initial information generation rule includes an initial execution flow and / or an initial calling interface; and the target information generation rule includes a target execution flow and / or a target calling interface. In the embodiment, the configuration information can be used to describe the initial execution flow required to be executed by the information processing model in the process of generating the reply information. The configuration information can also be used to describe the initial calling interface required to be called by the information processing model in the process of generating the reply information. The initial execution flow and the initial calling interface described above can be used to guide the information processing model to determine the target execution flow required to be executed in the process of generating the reply information. The initial execution flow and the initial calling interface described above can be used to guide the information processing model to determine the target calling interface required to be called in the process of generating the reply information. The initial information generation rule can include the initial execution flow and the initial calling interface. The target information generation rule can include the target execution flow and the target calling interface. As an optional implementation manner, the method further includes: creating parameter information corresponding to the reply information, wherein the parameter information is used to define key content in the reply information, and the configuration information and the parameter information are used to generate the prompt information of the information processing model. In the embodiment, the parameter information corresponding to the reply information can also be created, wherein the parameter information can be used to define the key content in the reply information, and the configuration information and the parameter information can be used to generate the prompt information of the information processing model. Optionally, the parameter configuration is an important part in the dialogue system. The parameter is a definition of key information in the dialogue question and answer link, and usually includes a name and description information. The large model can analyze the expression of the user and flexibly fill or change the parameter value when necessary. Optionally, the parameter information configured by the B-end operator is called. Through the above configuration information and the parameter information, the Prompt can be assembled. And the Prompt is sent to the large model Agent. oAs an optional implementation, the method further comprises: vectorizing the different configuration information to obtain different vectors; and storing the different vectors in a vector database. In this implementation, the configuration information can be vectorized to obtain a vector corresponding to the configuration information, and the vector obtained through the vectorization can be stored in the vector database. Optionally, after the B-side operation personnel configures the process information and the API plug-in information on the operation personnel management page, the information can be vectorized and saved in the vector database. The purpose of the method is to enable the computer to better understand and process the text information corresponding to the process information and the API plug-in information. The disclosure further provides another processing method of interactive information from a task-oriented dialogue scenario. FIG. 5 is a flowchart of another processing method of interactive information according to an embodiment of the disclosure. As shown in FIG. 5, the method can comprise the following steps: step S502, obtaining inquiry information, wherein the inquiry information is used to represent information to be replied in a task-oriented dialogue scenario. In the technical solution provided in the above step S502 of the disclosure, inquiry information to be replied sent by a user can be obtained. In this implementation, if the user has a demand for a dialogue robot to answer and reply related questions, the user can perform a corresponding input operation on an operation interface on a client device, through which the user can dialogue with the dialogue robot, to input user expressions that meet the user's query and questioning demand. When the operation interface detects that the inquiry information has been input, the inquiry information can be transmitted to a corresponding server through a network for corresponding analysis, to obtain reply information for replying to the inquiry information. Step S504, obtaining configuration information associated with the inquiry information, wherein the configuration information is used to describe initial information generation rules required to be met in a process in which a dialogue model generates reply information corresponding to the inquiry information, the initial information generation rules are used to guide the dialogue model to determine target information generation rules required to be met in a process in which the dialogue model generates the reply information, and the dialogue model is trained based on inquiry information samples in a task-oriented dialogue scenario and reply information samples in the task-oriented dialogue scenario. In the technical solution provided in the above step S504 of the disclosure, configuration information associated with the inquiry information can be determined. Optionally, configuration information with a high similarity to the inquiry information can be determined from a vector database as the configuration information associated with the inquiry information. Step S506, generating prompt information of the dialogue model based on the configuration information and pre-obtained parameter information associated with the inquiry information, wherein the parameter information is used to define key content in the reply information.In the technical solution provided in the foregoing step S506 of the present disclosure, after the configuration information associated with the query information is obtained, the prompt information of the information processing model can be generated based on the configuration information and the parameter information. Optionally, the parameter information configured by the B-end operator is called. By using the foregoing configuration information and the parameter information, the Prompt can be assembled. And the Prompt is sent to the large model Agent. oIn the technical solution provided in the step S508 of the present disclosure, after the prompt information of the information processing model is generated based on the configuration information and the parameter information, the prompt information can be used to guide the information processing model to generate the reply information corresponding to the inquiry information according to the initial information generation rule and the target information generation rule. Optionally, after the large model Agent receives the prompt, autonomous planning can be performed based on the configuration information and the parameter information, flexible jumping within the process can be realized, and the dynamic calling capability of the API can be realized. The above settings can enable the large model Agent to better understand the user's intention and more accurately schedule between the process and the API plug-in. By using the above large model Agent capability, a more fluent and natural human-like reply, that is, the reply information, can be generated. For example, after the information processing model generates the reply information corresponding to the current inquiry information, the reply information can be transmitted to the operation interface on the corresponding client device through the network and can be displayed in the form of a conversation with the user. For example, the inquiry information sent by the user can be displayed in the form of a bubble on the operation interface, and the reply information can also be displayed in the form of a bubble, but the directions of the two bubbles are different. The first bubble can point to the avatar displayed by the user in the operation interface. The second bubble can point to the avatar displayed by the dialogue robot in the operation interface. It should be noted that the display method of the above dialogue on the operation interface is only an example and is not limited specifically herein. Through the steps S502 to S508 of the present disclosure, the inquiry information is obtained, wherein the inquiry information is used to represent the information to be replied in the task-oriented dialogue scenario; the configuration information associated with the inquiry information is obtained, wherein the configuration information is used to describe the initial information generation rule required to be met by the dialogue model in the process of generating the reply information corresponding to the inquiry information, and the initial information generation rule is used to guide the dialogue model to determine the target information generation rule required to be met in the process of generating the reply information; the prompt information of the dialogue model is generated based on the configuration information and the parameter information associated with the inquiry information obtained in advance, wherein the parameter information is used to define the key content in the reply information; and the prompt information is used to guide the dialogue model to generate the reply information including the key content according to the initial information generation rule and the target information generation rule, thereby realizing the technical effect of improving the processing efficiency of the interactive information and solving the technical problem of low processing efficiency of the interactive information. The above method of the embodiment will be further introduced below.As an optional implementation, the query information and the reply information are multi-modal information, and types of the multi-modal information include at least one of the following: text information containing character information, video frame information containing frame image information, and audio information. Types of the reply information include at least one of the following: text information, image information, video information, and audio information. In this implementation, the query information and the reply information can be multi-modal information. The multi-modal information can include various types of information, and the information can be in the form of text, image, video, or voice. The text information can include text information containing character information, which can be a question, a request, or other forms of textual description proposed by the user. The video frame information refers to a single image frame in a video stream, and the image frame can contain rich visual information such as objects, scenes, actions, etc. The audio information refers to a sound signal, which can include voice conversation, music, environmental sound, etc. The dialogue system can reply to the user's question in the form of text, providing a textual answer or explanation. The dialogue system can generate an image containing visual content to reply to the user in the form of an image, such as generating a chart, visualizing structured data, etc. The dialogue system can reply to the user in the form of a video, such as demonstrating operation guidance or demonstration through a video. The dialogue system can reply to the user in the form of voice, providing a voice answer or guidance. In the present disclosure, by setting different types of query information and reply information, various forms such as text, image, video, and voice are covered. The dialogue system can reply in an appropriate form according to different types of input and needs of the user, to provide a richer and more diversified dialogue interaction experience. It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present disclosure, such as data for verification, are all information and data authorized by the user or fully authorized by all parties, and the collection, use, and processing of related data need to comply with relevant laws, regulations, and standards of relevant countries and regions, and provide corresponding operation entrances for the user to select authorization or refusal. According to the embodiments of the present disclosure, a dialogue system is also provided. It should be noted that the dialogue system of this embodiment can be used to execute the processing method of the interactive information of the embodiments of the present application. FIG. 6 is a schematic diagram of a dialogue system according to an embodiment of the present disclosure. As shown in FIG. 6, the dialogue system 600 can include a client 602 and a server 604. The client 602 is configured to obtain input information to be replied.The client is configured to generate prompt information of the information processing model based on the configuration information and pre-acquired parameter information associated with the input information, and generate reply information including key content using the prompt information, wherein the parameter information is used to define the key content in the reply information. In this embodiment, the input information to be replied can be acquired by the client 602. The client 602 can also generate prompt information of the information processing model based on the configuration information and parameter information corresponding to the reply information, and guide the information processing model to generate corresponding reply information according to the initial information generation rule and the target information generation rule using the prompt information, wherein the client can be a client device. Optionally, if a user has a demand for a dialogue robot to answer and reply related questions, the user can perform a corresponding input operation on an operation interface of the client 602_h capable of dialogue with the dialogue robot to input user expressions meeting the user's query and questioning demand. When the operation interface detects that there is input information that has been input, the input information can be transmitted to the corresponding server 604 through the network. The server 604 is configured to issue configuration information associated with the input information to the client, wherein the configuration information is used to describe the initial information generation rule required to be met by the information processing model in the process of generating the reply information corresponding to the input information, and the initial information generation rule is used to guide the information processing model to determine the target information generation rule required to be met in the process of generating the reply information. In this embodiment, the server 604 can issue configuration information associated with the input information to the client, and also generate corresponding prompt information based on the configuration information and pre-acquired parameter information associated with the input information. The prompt information is used to guide the information processing model to generate the reply information according to the initial information generation rule and the target information generation rule, and the reply information can be transmitted to the client 602 for display. Optionally, the server 604 can acquire the input information described above, and input the input information into the information processing model to process the input information. In the processing process, the initial information generation rule required to be met by the input information, i.e., the configuration information of the information processing model, can be determined. The prompt information guiding the information processing model to process the input information to generate the reply information can be generated based on the configuration information and the parameter information corresponding to the reply information. The prompt information is used to guide the information processing model to determine the target information generation rule required to be met in the process of generating the reply information, so that the reply information including key content can be generated according to the target information generation rule. Optionally, after the reply information is generated, the reply information can be transmitted to the client 602_h for display through the network.Further, the method is described as follows. As an optional implementation, the server comprises: an information configuration device configured to vectorize different configuration information to obtain vectors of the different configuration information, and store the vectors of the different configuration information in a vector database; and a client configured to vectorize input information to obtain a first vector, and send an information query request to the information configuration device, wherein the information query request comprises the first vector. The information configuration device is configured to, in response to the information query request, query, from the vector database, a second vector having a similarity to the first vector greater than a similarity threshold, and send configuration information corresponding to the second vector to the client. In this embodiment, the information configuration device for vectorizing configuration information can be included in the server. In the information configuration device, the configuration information can be vectorized to obtain corresponding vectors, and the vectors obtained by the vectorization can be stored in the vector database. Optionally, after the client vectorizes the input information to obtain the first vector, the information configuration device can be sent the information query request, so as to traverse the vectors in the vector database and find a vector meeting the requirement from a large number of vectors. The information query request can comprise the first vector. Optionally, after the information configuration device receives the information query request from the client, the vectors stored in the vector database can be traversed to find a vector having a similarity to the first vector in the information query request greater than a similarity threshold, and the vector is taken as the second vector. The second vector meeting the similarity threshold can be transmitted to the client. For example, when a user of the C end has a conversation, the expression of the user can also be vectorized to obtain the first vector. The first vector can be filled in the corresponding information query request, and the information query request can be used to call the vector database. After the vector database receives the information query request, a vector similarity algorithm can be used to calculate process information or API plug-in information having a similarity to the expression of the user higher than a similarity threshold, for example, the vector similarity algorithm can be used to determine five process information or API plug-in information close to the expression of the user, that is, the vector database can find the vectorized representation of the configuration information close to the expression of the user. The above process means that the conversation system can find process information or API plug-in close to the expression of the user, so as to quickly find possible relevant content from a large number of information.By the above dialogue system of the disclosure, the input information to be replied is acquired by the client 602; the configuration information associated with the input information is issued to the client by the server 604, wherein the configuration information is used to describe the initial information generation rule required to be met in the process of generating the reply information corresponding to the input information, the initial information generation rule is used to guide the information processing model to determine the target information generation rule required to be met in the process of generating the reply information, the client 602 generates the prompt information of the information processing model based on the configuration information and the parameter information associated with the input information acquired in advance, and generates the reply information including the key content by using the prompt information, wherein the parameter information is used to define the key content in the reply information, thereby realizing the technical effect of improving the processing efficiency of interactive information, and solving the technical problem of low processing efficiency of interactive information. At present, the related NLP technology has been the main application direction of the dialogue robot, and the technology has a wide use in the field of intelligent customer service and the like. For the task-oriented dialogue scene, the dialogue system usually adopts a pipeline architecture design, which includes multiple components, each component is responsible for executing different tasks, and the user utterance is executed in each component in turn and finally generates a reply. In the above architecture, different components can include natural language understanding (NLU) component, dialogue management component, task logic processing component and natural language generation (NLG) component and the like. The utterance input by the user is first understood and analyzed by the NLU component to extract the user intent and semantic information. Then, the dialogue management component can manage and control the dialogue flow according to the user intent and the context to determine the next operation. Then, the task logic processing component can call the related task logic and API according to the user demand to execute specific tasks. Finally, the NLG component can convert the result generated by the dialogue system into a natural language reply and return it to the user. The above pipeline architecture setting makes the dialogue system be able to perform multi-level processing and analysis on the user input, so as to more accurately understand the user demand and execute the corresponding task. In addition, each component can be optimized and customized according to the demand to improve the performance and accuracy of the dialogue system. In summary, the dialogue robot adopts the related NLP technology in the task-oriented dialogue scene, and through the pipeline architecture design, multiple components are included to be responsible for different tasks to realize the processing and reply generation of the user utterance. The above design makes the dialogue system be able to achieve wide commercialization landing results in the field of intelligent customer service and the like, and provide more intelligent and efficient services for users.However, the dialogue system of the above pipeline architecture relies heavily on B-end operation personnel configuration, and the operation personnel needs to be familiar with the task process, but also needs to learn complex product knowledge when configuring the canvas. The canvas configuration is complicated, the number of nodes is large, the node connection is intertwined and crossed, the maintainability is poor, and the reply based on pure rule configuration lacks fluency, is fixed and rigid. Therefore, there is still a technical problem of low processing efficiency of interactive information. The present disclosure provides a large model Agent task-based dialogue method, which solves the technical problem of low processing efficiency of interactive information, and is different from the related solution in which the B-end operation personnel configuration is complicated and the maintainability is poor, solving the technical problem of low processing efficiency of interactive information. In the embodiment of the present disclosure, an end-to-end large model task-based dialogue solution is provided, which processes NLU, DM and NLG by a large model, and integrates the capability of a large model API plug-in to greatly improve semantic understanding and dialogue flexibility. In terms of operation capability, the method provides a task flow-oriented simple canvas interaction scheme for platform users, which facilitates them to quickly configure the task flow into a dialogue flow. Since the embodiment of the present disclosure considers that the dialogue system is configured by operation personnel, which reduces the efficiency of the whole interactive information processing, a configuration information related to the task flow of the interactive information processing in the dialogue system can be defined, which is easy to learn and maintain, and can effectively avoid giving a large amount of interference information to the information processing model. Relying on the reasoning capability and instruction compliance of the information processing model, using the prompt information generated by the configuration information, the information processing model can determine the target information generation rule in the process of generating the reply information, that is, to perform autonomous planning of the flow, without defining the corresponding dialogue logic manually. Thus, the purpose of the information processing model to generate the reply information of the reply input information smoothly and naturally is achieved, thereby realizing the technical effect of improving the processing efficiency of interactive information, solving the technical problem of low processing efficiency of interactive information. The above method of the embodiment will be further introduced below. In the embodiment, NLP technology has always been the main application direction of dialogue robots, and the technology has a wide application in the field of intelligent customer service. FIG. 7 is a schematic diagram of a dialogue system of a pipeline architecture in a related technology, as shown in FIG. 7, for a task-based dialogue scene, the dialogue system usually adopts a pipeline architecture design, which includes multiple components, each component is responsible for performing different tasks, and user utterances are executed in each component in turn, and finally a reply is generated. In the above architecture, different components can include a natural language understanding component 702 (NLU component), a dialogue management component 703 (DM component), and a natural language generation component 704 (NLG component).The dialogue management component 703 can include a dialogue state tracking (DST) component 7031 and a dialogue policy component 7032, i.e., a DP component 7032. The user input utterance, i.e., the query information 701, is first understood and parsed by the NLU component 702 to extract the user intent and semantic information. Then, the DM component 703 can manage and control the dialogue flow according to the user intent and context to determine the next operation. Next, the task logic processing component can call the relevant task logic and API according to the user demand to perform a specific task. Finally, the NLG component 704 can convert the results generated by the dialogue system into a natural language reply, i.e., the reply information 705, and return it to the user. If the user's query information 701 is to inquire about the weather, the NLU component 702 analyzes the user's intent from the query information 701 to be the weather inquiry intent. The NLG component 704 can ask the user to query the city, for example, can issue the reply information "Please tell me which city you want to check the weather". Optionally, as shown in FIG. 7, the NLU component 702 is responsible for analyzing and understanding the user's input sentence, including extracting the user's intent, understanding semantic information, and extracting core parameters. Through natural language understanding technology, the NLU component can convert the user's utterance into structured data for subsequent processing of the dialogue system. The above components can usually be implemented using semantic parsing, entity recognition, part-of-speech tagging, etc. to ensure accurate understanding of user input. Optionally, as shown in FIG. 7, the DST component 7031 is responsible for tracking and maintaining the user's dialogue state, including the user's intent, context information, etc. The DP component 7032 is responsible for making decisions based on the user's intent and dialogue state based on the task flow to determine the behavior of the dialogue system in this round of dialogue. The above components can maintain the coherence of the dialogue and determine the next action of the dialogue system in combination with the task rules and strategies. Optionally, as shown in FIG. 7, the NLG component 704 is responsible for generating a natural language reply according to the behavior of the dialogue system. The output results of the dialogue system can be converted into natural language text and returned to the user. Through natural language generation technology, the NLG component can ensure that the reply of the dialogue system is natural and fluent, and conforms to the language expression habit. The above pipeline architecture design enables the dialogue system to perform multi-level processing and analysis on user input, thereby more accurately understanding user demand and performing corresponding tasks. In addition, each component can be optimized and customized according to demand to improve the performance and accuracy of the dialogue system.Overall, dialogue robots use relevant NLP techniques in task-oriented dialogue scenarios and are designed through a pipeline architecture, which includes multiple components responsible for different tasks to achieve user utterance processing and reply generation. The above design enables dialogue systems to achieve widespread commercialization in fields such as intelligent customer service and provide more intelligent and efficient services for users. In related technologies, task-oriented dialogue systems usually complete user tasks by combining multiple components. In actual task requirements, B-end operators need to describe task processes through configuration canvases, and components such as DP components and NLG components need to refer to task processes to generate results. B-end operators need to be able to describe task processes through configuration canvases. The above process means that visual tools or configuration interfaces can be used to design and define the task logic and processes of dialogue systems. The above task process description may include user intent, dialogue process, task rules, operation steps, etc. Through the above method, operators can intuitively describe and define the task logic that the dialogue system needs to perform without directly programming or system development. DP components need to refer to task processes to decide the behavior of the dialogue system in the dialogue. By referring to the task process, DP can make decisions based on task rules according to the user's intent and dialogue state. The above process includes determining the response of the dialogue system to the user's request, the next operation, and how to maintain the coherence of the dialogue. The NLG component needs to generate natural language replies according to the task process. By referring to the task process, NLG can ensure that the generated reply matches the task logic and guarantees the natural fluency of the reply. The above setting can make the dialogue system's reply to the user more consistent with the task requirements and rules. In related technologies, through the above design, operators can directly participate in the definition and description of task processes, and the dialogue management and natural language generation components of the dialogue system can make decisions and generate results according to the above task processes. The combination and cooperation of the above components enable task-oriented dialogue systems to better meet actual task requirements and provide more intelligent and personalized dialogue services. For example, FIG. 8 is a schematic diagram of a dialogue system generating a reply result process based on a pipeline architecture in a related technology. As shown in FIG. 8, based on the pipeline architecture, if the scenario is a weather query scenario, the user can input relevant input information to query the weather. The dialogue system can obtain the user's city information, call the weather query interface, and generate a weather query reply. However, in the above related technology, B-end operators are familiar with task processes, but configuration canvases need to learn complex product knowledge, increasing the learning cost. As the complexity of tasks increases, a large number of nodes and connections in the canvas make it difficult to maintain the canvas.The above process shows that the current configuration method is not user-friendly, and a more intuitive and simplified way is needed to describe the task flow. The NLU technology based on semantic understanding requires B-end operation personnel to configure a large amount of intention data, and it is inevitable to have intention conflicts, which affects the intention recognition effect. In addition, the intention understanding is only based on the intention configuration data, and does not care about the task flow, resulting in weak semantic generalization ability and being unable to infer the user's semantics well. The dialogue process has randomness, and it is difficult to predict the user's dialogue behavior. The dialogue management technology strongly depends on the process canvas configured by the B-end operation personnel, and cannot flexibly perform jumping, pulling back, and filling and modifying parameters. The above process shows that the current dialogue management technology needs a more intelligent and flexible way to handle the dialogue flow. The reply based on pure rule configuration lacks fluency, is fixed and rigid, and cannot provide a natural and humanized dialogue experience. The above process shows that the related natural language generation technology needs a more flexible and natural way to generate a reply. Moreover, when an error result is generated by a certain component, the subsequent components can process based on the error of the previous component, and it is also possible to generate an error result, which ultimately affects the final result. The above process shows that the current dialogue system architecture cannot effectively handle the error accumulation problem, and a more robust error handling mechanism is needed. Therefore, the above related technologies still have the technical problem of low processing efficiency of interactive information. In the embodiments of the present disclosure, an end-to-end large model task-based dialogue solution is provided, which has innovativeness, by processing NLU, DM and NLG by a large model, and integrating the ability of a large model API plug-in, to greatly improve semantic understanding and dialogue flexibility. In terms of operation capability, the solution provides a task flow-oriented minimalist canvas interaction solution for platform users, which facilitates them to quickly configure a task flow into a dialogue flow. FIG. 9 is a schematic diagram of a logical architecture of an end-to-end large model task-based dialogue according to an embodiment of the present disclosure, as shown in FIG. 9, the logical architecture can include an access layer, a data storage (Data Storage, referred to as DS) engine layer and a model layer. The access layer can include a dialogue scheduling central control 901. The DS engine layer can include NLU preprocessing 902, DM 903 and NLG 904. The NLU preprocessing 902 can include candidate processing 9021, NLU-LLM 9022 and API+flow recall 9023. The DM 903 can include context processing 9031, priority strategy 9032 and API execution 9033. The NLG 904 can include intent classification 9051, flow execution 9052 and Plugin 9052.The logical architecture can also include operation capabilities 906, which can include process building 9061, process testing 9062, process publishing 9063, API plug-in management 9064, and parameter management 9065. By handing over NLU, DM, and NLG tasks to large models, the dialogue system can achieve end-to-end dialogue processing, and through large-scale model training, the large model can better understand the user's intent and semantics, thereby improving the semantic understanding ability and dialogue flexibility of the dialogue system. The above design can enhance the overall performance of the dialogue system. The ability to integrate large model API plug-ins can add rich external services and data sources to the dialogue system, enabling the dialogue system to more flexibly call external APIs, such as map services, weather services, payment services, etc., thereby improving the functionality and practicality of the dialogue system. The platform provides a task flow-oriented minimalist canvas interaction solution for platform users, enabling users to quickly configure task flows into dialogue flows. The above design can reduce the learning cost of platform users, making it more intuitive and simplified to configure the dialogue system, and improving user friendliness. Through the processing of large models, the dialogue system can better handle complex semantic understanding, and the flexibility of large models can make the dialogue system more adaptable to different task scenarios and user needs, providing more accurate, natural, and personalized dialogue experiences. In summary, the above end-to-end large model task-based dialogue solution integrates large model processing of all tasks, integrates the ability of large model API plug-ins, and provides a task flow-oriented minimalist canvas interaction solution, which can improve the performance, flexibility, and user friendliness of the dialogue system, bringing new opportunities for the development of task-based dialogue systems. FIG. 10 is a flowchart of an end-to-end large model task-based dialogue processing according to an embodiment of the present disclosure, as shown in FIG. 10, the processing process can include process configuration information 1001, API plug-in configuration information 1002, parameter configuration information 1003, vector database 1004, process / API plug-in coarse recall 1005, assemble Prompt 1006, API 1007, plug-in call 1008, parse result 1009, and large model Agent 1010. The following steps can be included: Step S11, configuring process configuration information, API plug-in configuration information, and parameter configuration information by B-end operation personnel. In this embodiment, the B-end operation personnel can configure the process configuration information, API plug-in configuration information, and parameter configuration information. Optionally, the B-end operation personnel can configure the process information through the management page. In addition to the name and description, the most important thing in the process is the canvas information.The operation personnel can describe the task flow by using the nodes and the lines in the canvas by combing the task flow. The nodes and the lines in the canvas flow can be compiled into the Mermaid syntax structure by the dialogue system and saved for processing by the large model agent. In the related technology, due to the complexity of the application dialogue itself, new nodes and lines may need to be added in the canvas to support. As the task demand and the dialogue scene continue to expand, the canvas configuration mode may not be able to cope with the complexity well, resulting in that the canvas configuration is too bloated and difficult to manage. The dialogue itself has randomness and needs to support flexible jumps. The dialogue management component may not be able to completely cover the above flexibility, resulting in the need to constantly add lines to solve the jump problem in the dialogue flow, so that the canvas configuration becomes complex and difficult to maintain. The NLU may have limitations, so that special processing logic needs to be written in the node for secondary processing. The above cases may lead to the logic of the node being complicated, increasing the complexity and difficulty of the configuration. In summary, the related canvas flow configuration is complex and difficult to maintain mainly due to the complexity of the dialogue, the randomness and flexible jump demand of the dialogue, and the limitations of the related NLU. In view of the above problems, a more intelligent, flexible and easy-to-maintain dialogue system design and configuration mode needs to be sought. However, in the embodiment of the disclosure, the canvas design scheme only needs to define the core task flow, relies on the inference ability and instruction compliance of the large model, and does not need to define all the dialogue logic artificially, so the nodes and edges on the canvas are relatively simple. The edges can directly edit the dialogue and similar questions, which are used for the large model to plan the flow. In addition to supporting direct editing of replies, the nodes can also call external API plugins and parameter collection, and at the same time, in the advanced model, the instructions can be directly described in a natural language manner, and the large model can follow the instructions to guide the dialogue. The canvas design scheme in the embodiment of the disclosure does not need to define all the dialogue logic artificially, but only needs to define the core task flow. Relying on the inference ability and instruction compliance of the large model, the above design can greatly simplify the configuration of the nodes and edges on the canvas, so that the configuration is more simple and intuitive. In the above design scheme, the nodes and edges on the canvas are very simple, the edges can directly edit the dialogue and similar questions, which are used for the large model to plan the flow. The above process makes the configuration of the task flow more intuitive and flexible. In addition to supporting direct editing of replies, the nodes can also call external API plugins and parameter collection. The above design makes the configuration of the nodes more flexible and can meet diversified task demands. In the advanced model, the nodes also support directly describing the instructions in a natural language manner, and the large model can follow the instructions to guide the dialogue.The above design makes the configuration of dialogue logic more intelligent and flexible, and conforms to the intelligent reasoning and instruction following capability of the large model. In general, the above canvas design scheme greatly simplifies the configuration process of the dialogue system, improves the intuitiveness and flexibility of the configuration, and brings new opportunities for the intelligentization and personalization of the dialogue system, by means of the reasoning capability of the large model, through simple node and edge configuration, support for natural language description instructions and the like. For example, FIG. 11 is a schematic diagram of a flow canvas of an electronic toll collection system transaction according to an embodiment of the present disclosure, as shown in FIG. 11, the flow canvas of the ETC transaction can first determine whether the user's intention is to purchase an ETC device or to activate an ETC. If the user's intention is to purchase an ETC device, the user can be introduced to the device type and recommended. It can be tried to persuade the user to select a card-free ETC. If the user has the intention of how to purchase, the purchase link can be sent. If the purchase link is sent, it can also be asked whether there are other needs. If the user's intention is to activate the ETC, the API can be called to query the order status. If the order status is a valid order, the API can be called to verify the user's identity. If the user's identity verification is successful, the user can be guided to complete the activation according to the activation instruction. If the activation is successful, the user can be congratulated on the successful activation, and thanked and goodbye. For another example, FIG. 12 is a schematic diagram of calling an external application programming interface plug-in according to an embodiment of the present disclosure, as shown in FIG. 12, the external API plug-in can be called. If the external API plug-in to be called needs to call the API to query the order status, the external API plug-in to be called at this time can be determined as the API calling plug-in to query the order status at the node name in the page of editing the node. And the normal mode or the advanced mode can be selected at the page of executing the action, and the definition of the API plug-in: getOrederStatus can be input to ask the license plate number and call the API plug-in information. As an optional example, FIG. 13 is a schematic diagram of a complex instruction in an advanced mode according to an embodiment of the present disclosure, as shown in FIG. 13, in the process of editing the device type introduction node, if the advanced mode is selected, in this case, the complex instruction in the advanced mode can be to perform the following operations in steps: first ask the user which type of ETC device is needed, including card-type ETC device and card-free ETC device.After the user selects the category, the corresponding device is introduced: 1. The user needs to insert the ETC card into the on-board card reader, complete the deduction through the on-board card reader, and the corresponding price is 68 yuan; 2. The card-free ETC is directly integrated into the vehicle's on-board unit (OBU), and the automatic deduction is achieved by interacting with the highway toll station. The card-free ETC is convenient and fast. Through the above operation, it can be seen that the user only needs to describe the task flow in the canvas and appropriately call external API plugins to complete the configuration of the flow canvas. The learning cost of the user is effectively reduced, and the ease of use of the canvas flow is improved. In the present embodiment, the above setting mode enables the user to only describe the task flow in the canvas and call external API plugins when necessary to complete the configuration of the flow canvas. The above design effectively reduces the learning cost of the user and improves the ease of use of the canvas flow. The user can describe the task flow in an intuitive way without understanding complex technical details or syntax rules, thereby reducing the learning cost of configuration. Through simple operation, the user can define the nodes, edges and external API calls in the task flow to complete the configuration of the entire dialogue system. At the same time, by calling external API plugins, the user can conveniently integrate external services and data sources, thereby enriching the functions of the dialogue system. The above design enables the user to more flexibly customize and configure the dialogue system to meet the needs of different task scenarios. In general, the above setting mode can effectively reduce the learning cost of the user and improve the ease of use of the canvas flow by simplifying the configuration process and integrating external API plugins, thereby providing a more intuitive, flexible and intelligent configuration mode for the user. In the embodiment, the B-end operator can configure API plugin information on the management page. The API plugin information mainly includes name, description and input / output parameter description. After the API plugin information is given to the large model by Prompt, the large model plans autonomously based on the API plugin name and description, calls the API plugin to obtain the result when necessary, and continues to plan based on the result to generate the final reply. The above setting allows the operator to flexibly configure the API plugin information, providing more external data and service support for the large model. The large model can autonomously plan and decide according to the API plugin information, call the API plugin to obtain the required result when necessary, and then use the result for the final dialogue reply generation. Through API plugin configuration, the dialogue system can more intelligently utilize external resources to provide more rich and personalized dialogue services.The above design can effectively improve the flexibility and intelligence of the dialogue system, and provide users with a more expected dialogue experience. For example, FIG. 14 is a schematic diagram of an application programming interface plug-in of an electronic toll collection system transaction according to an embodiment of the present disclosure. As shown in FIG. 14, in the ETC transaction scenario, if the service name is getOrederStatus and the service description is to query the ETC order status according to the license plate number, interface configuration can be performed at the API configuration in the page, and the parameter name, type, value source, and whether it is required of the corresponding input parameter can be input. The parameter name, type, and parameter interpretation of the output parameter can also be added at the new parameter. In this embodiment, in addition to the flow and API, another important configuration item is the parameter. The parameter is a definition of key information in the dialogue question and answer link. The parameter can include the name and description information. The large model can analyze the user's expression and flexibly fill or change the parameter value. For example, FIG. 15 is a schematic diagram of a parameter of an electronic toll collection system transaction according to an embodiment of the present disclosure. As shown in FIG. 15, in the ETC transaction scenario, if parameter configuration is required, the name and description of the parameter can be added at the new parameter in the page, for example, if the name of the new parameter is ETC and the description is that the card insertion includes card insertion type ETC and card-free type ETC. Step S12, vectorizing the flow configuration information and API plug-in configuration information and storing them in the vector database. In this embodiment, after the B-end operator manages the page configuration of the flow information and the API plug-in information, the above information can be vectorized and saved in the vector database. The purpose of the above method is to enable the computer to better understand and process the text information corresponding to the flow information and the API plug-in information. Step S13, sending a user expression and triggering flow / API plug-in coarse recall. In this embodiment, the user's expression can be received, and flow / API plug-in coarse recall can be triggered. Optionally, using the vector database to perform coarse recall of the flow / API plug-in can largely discard other flow / API plug-ins irrelevant to the user's expression, avoiding a large amount of irrelevant information from being given to the large model. The above process reduces the model prediction cost, because the large model does not need to process a large amount of irrelevant information, and also improves the model prediction accuracy, avoiding the model from mistakenly recalling irrelevant flows or API plug-ins. In general, through the vector database, coarse recall of the flow / API plug-in can effectively reduce the interference of irrelevant information, improve the efficiency and prediction accuracy of the model, and thus provide more accurate and fast dialogue services for users.Optionally, after the B-side operator configures the process and API plug-in information on the management page, the process information and API plug-in information can be vectorized and saved in the vector database. The user expression of the C-side can also be vectorized to call the vector database, and the vector database internally calculates the 5 processes or API plug-in information closest to the user expression through a vector similarity algorithm. Using the vector database for coarse recall of processes / API plug-ins can largely discard other processes / API plug-ins irrelevant to the user expression, avoiding a large amount of irrelevant information from being fed into the model, which reduces the model prediction cost and improves the model prediction accuracy, avoiding the model from mistakenly recalling irrelevant processes or API plug-ins. Step S14, vectorizing the user expression. In this embodiment, the received user expression can be vectorized. Optionally, after the user expression is vectorized, the dialogue system can calculate the similarity between the first vector and the vectors in the vector set corresponding to each scenario in the vector database. Common similarity calculation methods include cosine similarity, Euclidean distance, etc. The above calculation methods can measure the similarity between two vectors. Step S15, assembling Prompt using three types of configuration information. In this embodiment, Prompt can be assembled. Optionally, in addition to the above process and API information, the parameter information configured by the operator can also be applied in the above components, and the above three types of information are combined to assemble Prompt to call the large model. Optionally, the Prompt structure can include task description, which is used to describe the role that the large model Agent needs to play, the task to be performed, the definition of input and output formats, and examples. It can also include process information, which can be obtained by calculating the process information from the vector database. The internal core task process is described by the operator through the canvas, and is represented using the Mermaid syntax structure. The large model can refer to this task process for planning and guiding the progress of the dialogue. It can also include API plug-in information, which is obtained by calculating the process information from the vector database. The internal API plug-in information is configured by the operator, and the large model can select to call the API when necessary and plan again based on the API result. It can also include parameter information, which is the key information in the dialogue process configured by the operator. The large model can extract the parameter information in combination with the context and understand based on the parameter information when necessary to complete the dialogue task. It can also include environment information, which can include current time and other environment information, used to assist the identification and planning of the large model. It can also include context information, which includes the multi-round user expression and the dialogue system reply of the multi-round conversation information of the C-side user.Optionally, after the large model Agent receives the Prompt, based on the process / API information, combined with the context state, it can realize flexible jumping within the process and dynamic calling ability of API. For example, in the ETC transaction, the following Prompt can be obtained: 1<

[0004] sys tem you are an intelligent assistant, now need to help the user ETC activation and consumption query VI specific requirements by answering the user's phone:

[0005] 1. Please remember that you are a public servant of a company, and answer the following task flow and API information questions intelligently.

[0006] 2. If the user asks a question beyond the following task flow and API, please answer that it is not clear.

[0007] 3. If the user's answer involves multiple possible intentions in the task flow or API information, please ask the user to clarify their specific intention.

[0008] 4. If the user has not completed the task, please guide them to complete the current task. If the user has completed the task or says goodbye, please politely bid farewell and end the conversation.

[0009] 5. When the user provides parameter information, please ensure that it is determined according to the specified format, including the correctness of the format and the completeness of the content. If the parameter does not meet the format requirements, please remind the user to modify it until it meets the requirements.

[0010] 6. You are a professional task officer with high response diversity and content that will not be repeated.

[0011] 7. The response format is: Thought: xxx

[0012] Response: xxx o Among them, Thought includes four pieces of information: [Step: Analyze the last node] last_node=xxx, [Step: Analyze the current intention] intent=xxx, [Step: Analyze the next node] current_node=xxx, [Step: Determine the next plan] I need xxx next.

[0013] 8. Thought represents the thinking process in the middle, Response is the content returned to the user, the nodes in the flowchart corresponding to the last_node and current_node in Thought represent the progress of the user handling the task, and the edge in the flowchart corresponding to the current intent represents the user's intention.

[0014] 9. The current intent and node are not within the scope of the flowchart and API information description, and are represented by UNK. The content of [Step: Determine the follow-up plan] refers to the textual description in the current node current_node. At the same time, clarify which case when asking: When multiple intentions are expressed, use AND to connect. The above two connection methods are also applicable to nodes.

[0015] 11. When the current node current_node is an API node, Thought needs to add three additional information after [Step: Analyze the next node]: [Step: Specify the required parameters of API] The mandatory parameters of xxx are: xxx, [Step: Extract the API parameter value] From the above, the xxx parameter and its value can be known: x=xxx.

[0016] 12. The following is a complete example of Thought, do not output [Step: xxx], and output according to the following format: Thought; last_node=C, current_intent=ETC consumption query, current_node=C, mandatory parameters of getETCRecord: license plate, date, no optional parameters, from the above, the parameters and values of getETCRecord can be known: license plate=ABG0375, getETCRecord still lacks the mandatory parameter: date.

[0017] 13. Your ability is based on the following task flow and API information:

[0018] ETC activation and consumption query vl processing flow flowchart TD

[0019] BQ000 (Start) - ETC Activation - > BQ001 > Current Target: Call API to query order status. Target Description: Based on the definition of getOrderStatus, query the license plate number and call the API. If no valid order is found, ask the user to confirm the license plate number information.

[0020] BQ000 (Start) - ETC Device Purchase - > BQ002 [Current Goal: Device Type Introduction Goal Description: Perform the following operations step by step: First, ask the user which type of ETC device they need. Currently, there are two types: card-insertion ETC and cardless ETC. After the user selects a category, introduce the corresponding device:]

[0021] 1. For the card-type ETC, the user needs to insert the ETC card into the vehicle's card reader. The system communicates with the toll station through the antenna on the vehicle's windshield to automatically complete the payment. The corresponding price is 68 yuan.

[0022] 2. Cardless ETC is directly integrated into the vehicle's OBU (On-Board Unit), eliminating the need for users to insert a card. It interacts with the toll station via radio waves to achieve automatic payment. Cardless ETC is convenient and fast, and its use is less affected by card issues. The corresponding price is 98 yuan.

[0023] BQ001 - Valid order found - BQ003 - Query parameters and call according to API definition: authent icateUser.

[0024] BQ002 — Device Recommendation 1 > BQ004 [Current Objective: To persuade users to choose the cardless ETC. Objective Description: To persuade users to choose the cardless ETC, whose advantages include: Convenience: Users do not need to insert or remove a card, reducing inconvenience and passage delays caused by forgetting to insert or remove the card. Stability: Due to the lack of frequent physical contact, cardless ETC devices are generally more durable than card-insertion ETC devices, reducing the failure rate. Security: Cardless ETC reduces the time vehicles spend at toll booths, reducing the risk of traffic congestion and accidents. Although the price is higher, it is more cost-effective.]

[0025] BQ003 —Verification Successful > BQ005 [Current Goal: Guide the user to complete activation based on activation knowledge. Goal Description: Activation Method: Attach the device to the right side of the rearview mirror on the windshield, turn on the device's Bluetooth via the Bluetooth button, connect the phone's Bluetooth, open the E-Highway APP, follow the prompts to complete device activation, and finally confirm the device status. If the device displays the billing card, the device is successfully activated.]

[0026] BQ004 - How to Purchase --> BQ006 - Based on the API: sentPurchaseLink definition, query parameters and call...

[0027] BQ005—Activation Successful—>BQ007 [Congratulations on successful activation, thank you and goodbye].

[0028] BQ006—Sent Complete—>BQ008 [Information that the purchase link has been sent, please check! Inquire if there are any other needs]. In the process—the middle indicates user intent or redirection conditions. Some user intents and their example statements are: "TC activation": ["Cannot activate", "According to the instructions, activation failed", "I just bought your ETC, and I activated it according to the instructions, but it still doesn't work."]

[0029] The API information is as follows:

[0030] API: authent icateUser Description: Verifies identity information based on the last 6 digits of the ID card number. Input parameters: ("id_nuniber_last": ("type": "string", "description": "last 6 digits of ID card number")

[0031] Six digits, usually six digits, but could also be five digits and one digit. Z X Z Output parameters: ("result": ("type": "string", description": "ID card verification result, successful verification result is, success") 7 Failure is defined as... 7 " ")).

[0032] API: sentPurchaseLink Description: Sends a purchase link according to ETC type. Input parameters: ("etc. type": ("type": "string", "description": "ETC type, including card-type ETC and non-card-type ETC", "required": true}) Output parameters: ("status": ("type": "string", "description": "Link sending result; success message indicates successful sending") 7 Failure is defined as... 7" "))。

[0033] API: getOrderStatus Description: Query ETC order status according to license plate number Input parameters: ("plan.no": ("type": "string", "description": "License plate number, 7 or 8 characters, the first character is the abbreviation of the province, followed by numbers and letters", "required": true}) Output parameters: ("order-status": ("type": "string", "description": "Order status, Remember the current time is: 2024-02-24 18: 07: 31 Saturday <[im_end]>

[0034] user

[0035] ETC device purchase <| im_end|>

[0036] assistant

[0037] Thought: last_node = UNK, intent =. In step S16, the API call parameter is sent. In this embodiment, the parsing result 1009 can send the API call parameter. In step S17, the API plugin of the API is called. In this embodiment, the API call parameter can be transmitted to the API 1007 to call the corresponding API plugin. In step S18, the API plugin result is sent. In this embodiment, the API 1007 can be controlled to send the API plugin result. In step S19, the Prompt is sent. In this embodiment, after the assembly of the Prompt is completed, the Prompt can be sent to the large model Agent 1010. Optionally, the large model can effectively enhance the semantic understanding ability, can more deeply understand the user's intent, and more accurately schedule between processes and API plugins. In addition, since the dialogue process itself is unpredictable, by understanding the complete task flow through the large model, flexible jumping within and between processes can be realized. At the same time, multi-intent expression of the user is also supported, and the large model can intelligently clarify the intent or simultaneously execute multiple processes. Finally, the fixed and rigid reply is a common problem in related dialogue systems, and by using the ability of the large model, a natural and smooth human-like reply can be generated. In step S20, the reply information is fed back to the C-end user. In this embodiment, the reply information generated by the large model Agent 1010 can be fed back to the corresponding C-end user. In the embodiment of the present disclosure, the proposed end-to-end task-oriented dialogue NLU+DM+NLG solution has the following two advantages: first, by using the planning and memory capabilities of the large model, based on process reasoning, the flexibility and personification of the dialogue are greatly improved. The problems of rigid DM and non-personification NLG in the dialogue process are well solved; second, in the canvas design, only the task flow definition needs to be interactively designed, and the core task flow is defined. The large model introduced in the case can automatically realize the process jump and abnormal query answer generation according to the context. The canvas is easy to use and maintain. For example, a high-speed information query system can be used as a dialogue scenario. The high-speed information query system can include road condition, toll station and service area query functions. The operation personnel help users understand the high-speed situation in real time through the task flow. FIG. 16 is a schematic diagram of a local high-speed information query system canvas in a related technology, as shown in FIG. 16, the related canvas configuration is complex, the number of nodes is large, and the node connection is also entangled and crossed, which is difficult to maintain.However, FIG. 17 is a schematic diagram of an end-to-end flow canvas of a high-speed information query system according to an embodiment of the present disclosure. As shown in FIG. 17, the canvas configuration is simple and the flow is clear. The canvas is constructed purely based on task flow, and the learning and configuration cost is extremely low. For example, the user can be greeted first to inquire about the user's needs. If the user's needs are to query the service area or parking area name, the service area or parking area name can be collected. If it is a parking area, it can be determined whether it is a service area or a parking area. If it is a service area with the same name, it can be determined that the service area is on a highway. If it is not a service area with the same name and a parking area, the service area information can be obtained by calling an API. If the user's needs are to query the road condition, the road segment information can be collected. If it is a road segment, the highway name can be determined. If it is a highway, the road segment name can be determined. If it is a letter and a number, the highway name can be determined according to the road code. If it is a highway and a road segment name, the highway condition information can be queried by calling an API. If the user's needs are to query the toll station, the toll station name can be collected. It can be determined whether it is an airport or an inquiry telephone, invoice, etc. If it is an airport, the airport toll station can be determined. If it is an inquiry telephone, invoice, etc., the consultation question reply can be performed. If it is a toll station, the toll station information can be queried by calling an API. For example, FIG. 18 is a schematic diagram of the effect of a dialogue question and answer in a high-speed information query system according to an embodiment of the present disclosure. As shown in FIG. 18, if the high-speed information query system is a dialogue scenario, in this scenario, if the user inputs the inquiry information "Hello, is the Rizhao toll station closed? I want to check the road condition information of the G1511 Rizhao section", the dialogue robot can reply "The Rizhao toll station is currently temporarily closed in both directions. For the specific opening time, please follow the Shandong Expressway Wisdom Travel WeChat applet to subscribe to the toll station traffic status information. The G1511 Rizhao section is currently open to traffic. If you need further assistance, please feel free to consult at any time." If the user inputs the inquiry information "When will it be open" based on the reply of the above dialogue robot, the dialogue robot can reply "I'm sorry, I don't have specific information about the opening time of the toll station. This is usually determined by the traffic police department. Please follow the Shandong Expressway Wisdom Travel WeChat applet to subscribe to the toll station traffic status information to get the latest opening time. If you have other questions, I will do my best to answer them." If the user inputs the inquiry information "What is the phone number" based on the reply of the above dialogue robot, the dialogue robot can reply "I'm sorry, I don't have the toll station phone number information. You can try to dial the local area code plus 96659. This number is usually the unified service hotline of the Shandong Expressway service area or toll station, and a staff member may be able to provide you with further assistance."If you need other help, I will try to answer you. Alternatively, as shown in FIG. 18, in the first round: the user's expression involves multiple intents, the large model agent understands multiple intents, planning based on the flowchart, collecting three parameters: Rizhao toll station, G1511 highway, Rizhao section, and successively executing two APIs of toll station information query and highway section query, and finally combining the results of the two APIs to generate a smooth and natural reply. It can be seen that the large model agent has strong semantic understanding ability, strong autonomous planning ability, and the generated reply is very smooth and natural; in the second round: the user's expression involves flow jump, and the large model agent flexibly jumps to the consultation question reply node based on the user's semantics and the flowchart. The flexible jumping ability of the large model agent based on the node avoids complex wiring in the canvas; in the third round: the user is still asking for toll station information, and it can be seen that the large model agent understands multiple complex instructions in the node, understands and follows the correct instructions to reply. In addition, the rich instruction target configuration in the node can greatly reduce the number of nodes in the canvas. FIG. 19 is a schematic diagram of reply content in a high-speed information query system according to an embodiment of the present disclosure. As shown in FIG. 19, in the scene of the high-speed information query system, in the consultation question reply, the robot can combine the reply content and the dialogue context to generate the following reply content: if the user asks when the toll station will be opened, reply that the toll station closure is decided by the traffic police department, and pay attention to the Shandong Express Intelligent Travel WeChat mini-program to subscribe to the toll station passage state information; if the user asks for the phone number of a certain toll station, reply that it is recommended to dial "area code plus 96659" for targeted artificial consultation. If the user asks to query the distance from the toll station, reply to use the WeChat search "Shandong Express Intelligent Travel" mini-program, click on the toll station, input the toll station name, and query the distance from the toll station. If the user asks whether the toll station can issue electronic invoices, reply that all 465 toll stations under the jurisdiction of Shandong Express have implemented paper-to-electricity, and the specific method of issuing is to dial 96659 for artificial consultation. For another example, the photo developing task system can be taken as a dialogue scene, and by collecting the user's requirements for developing photos, such as number of photos, size, etc., the user is helped to develop photos and place an order for home delivery. FIG. 20 is a schematic diagram of an end-to-end flowchart canvas of a photo developing task system according to an embodiment of the present disclosure. As shown in FIG. 20, if it is detected that the user's query requirement is to develop photos or develop photos, it can be determined whether to recommend the size or explain what plastic sealing is. If the size needs to be recommended, the size can be recommended to the user. If it is necessary to answer what plastic sealing is, the user can be introduced to plastic sealing.The size can be collected, and the number of collections can be determined for 3 inches to 7 inches. If it is for other sizes, the user can be replied that it is not supported. If the number of collections exceeds 500, it can be replied to transfer to manual. If the number is within 500, the parameter calling API can be inquired to provide a link to the user. If the user indicates yes, it can be directly operated by the brand, and if the user indicates later, it can be flexibly replied to the user in multiple times. If the user indicates that it is a bit expensive, the user can be replied that the current price is also a limited-time special offer, and it is recommended to order as soon as possible. If the user inquires whether there is a discount, the user can be replied with a discount. FIG. 21 is a schematic diagram of a dialog question and answer effect of a photo developing task system according to an embodiment of the present disclosure. As shown in FIG. 21, in the scene of the photo developing task system, if the user inputs the inquiry information of “I want to develop photos”, at this time the dialog robot can reply “Rizhao toll station is temporarily closed at present, please pay attention to Shandong Highway Wisdom Travel WeChat mini-program to subscribe to toll station traffic state information for specific opening time, G1511 Rizhao section is currently open to traffic, if you need other help, please feel free to consult at any time.” If the user inputs the inquiry information of “10, what size can you recommend” based on the reply of the above dialog robot, at this time the dialog robot can reply “Hello, the recommended size of the photo is 5 inches, which is 【8.9*12.7cm】, and it is a common size when using film printing in the early years, the ratio is 10:7. Do you need a 5-inch photo?” If the user inputs the inquiry information of “that's 5 inches, I'll tell you tomorrow, I'm a bit busy today” based on the reply of the above dialog robot, at this time the dialog robot can reply “OK, you are busy first, the photo is 5 inches, tell me your specific needs tomorrow, if you need to order now, you can upload the photo when you are convenient. Or you can collect the order or add it to the cart, welcome to consult at any time. Wish you a smooth work today”. Optionally, as shown in FIG. 21, the first round: the user explicitly indicates that he wants to develop photos, the large model Agent understands the intention, based on the task flow configuration, it is clear that the number of photos is an important parameter, so the number of photos is inquired. The second round: after the user replies the number of photos, the user expects to recommend the size, the large model Agent flexibly jumps to the recommended size node based on the canvas configuration, and directly asks the user whether the recommended size is needed after recommending the size. The large model Agent has strong process understanding and reply generation capability, and clearly that the size is an important parameter, directly confirms with the user whether to use the recommended size combined with the currently recommended size.In the third round, after the photo size is collected, the large model initiates an API plug-in call and generates an order, but when the user explicitly indicates that he / she is busy, the dialogue system does not continue to disturb the user, but instead ends the conversation with the user in a friendly manner, and it can be determined that the order is valid for a long time. The large model agent can generate friendly and natural anthropomorphic replies in combination with the context, which is very important in the field of customer service scenarios. In the embodiment of the present disclosure, if it is necessary to reply to the input information to be replied, the input information can be obtained, and the input information can be input into the information processing model to process the input information. In the processing process, the initial information generation rule required to be met by the input information can be determined, that is, the configuration information of the information processing model. Based on the configuration information and the corresponding parameter information obtained in advance, the prompt information for guiding the information processing model to process the input information to generate the reply information can be generated. Through the prompt information, the target information generation rule required to be met by the information processing model in the process of generating the reply information is determined. Therefore, the reply information containing the key content can be generated according to the target information generation rule. Since the dialogue system configuration requires too much participation of the operation personnel, which can cause the efficiency of the whole interactive information processing to be reduced, the embodiment of the present disclosure can define a configuration information related to the task flow of the interactive information processing in the dialogue system, which is easy to master and maintain, and can effectively avoid giving a large amount of interference information to the information processing model. Based on this, the embodiment of the present disclosure relies on the reasoning ability and instruction compliance of the information processing model, and uses the prompt information generated by the configuration information, so that the information processing model can determine the target information generation rule in the process of generating the reply information, that is, to perform flow autonomous planning, without defining the corresponding dialogue logic artificially, thereby achieving the purpose that the information processing model can generate the reply information of the input information smoothly and naturally, and further achieving the technical effect of improving the processing efficiency of the interactive information. The technical problem of low processing efficiency of interactive information is solved. According to the embodiment of the present disclosure, an interactive information processing apparatus for implementing the interactive information processing method shown in FIG. 2 is also provided. FIG. 22 is a schematic diagram of an interactive information processing apparatus according to an embodiment of the present disclosure, as shown in FIG. 22, the interactive information processing apparatus 2200 can include a first obtaining component 2202, a second obtaining component 2204, a first generating component 2206, and a second generating component 2208.The first obtaining component 2202 is configured to obtain input information to be replied; the second obtaining component 2204 is configured to obtain configuration information associated with the input information, wherein the configuration information is used to describe an initial information generation rule required to be met in a process in which an information processing model generates reply information corresponding to the input information, and the initial information generation rule is used to guide the information processing model to determine a target information generation rule required to be met in the process of generating the reply information; the first generating component 2206 is configured to generate prompt information of the information processing model based on the configuration information and pre-obtained parameter information associated with the input information, wherein the parameter information is used to define key content in the reply information; and the second generating component 2208 is configured to generate the reply information including the key content by using the prompt information. Here, the first obtaining component 2202, the second obtaining component 2204, the first generating component 2206 and the second generating component 2208 correspond to steps S202 to S208 in the above embodiment, and the four components have the same instances and application scenarios as the corresponding steps, but are not limited to the content disclosed in the above embodiment. It should be noted that the above components can be hardware components or software components stored in a memory (for example, the memory 2604) and processed by one or more processors (for example, the processors 2602a, 2602b,..., 2602n), or the above components can be run in the computer terminal A provided in the embodiment of the present disclosure as a part of the apparatus. According to the embodiment of the present disclosure, an interactive information processing apparatus for implementing the interactive information processing method shown in FIG. 3 is further provided. FIG. 23 is a schematic diagram of another interactive information processing apparatus according to an embodiment of the present disclosure. As shown in FIG. 23, the interactive information processing apparatus 2300 can include a third obtaining component 2302, a first determining component 2304, a fourth obtaining component 2306, a third generating component 2308 and a fourth generating component 2310.The third obtaining component 2302 is configured to obtain input information to be replied; the first determining component 2304 is configured to determine a scenario to which the input information belongs; the fourth obtaining component 2306 is configured to obtain configuration information associated with the input information in the scenario, wherein the configuration information is used to describe an initial information generation rule required to be met in a process in which an information processing model generates reply information corresponding to the input information, and the initial information generation rule is used to guide the information processing model to determine a target information generation rule required to be met in the process of generating the reply information; the third generating component 2308 is configured to generate prompt information of the information processing model based on the configuration information and pre-obtained parameter information associated with the input information, wherein the parameter information is used to define key content in the reply information; and the fourth generating component 2310 is configured to generate, by using the prompt information, the reply information that meets the scenario and includes the key content. It should be noted that the third obtaining component 2302, the first determining component 2304, the fourth obtaining component 2306, the third generating component 2308 and the fourth generating component 2310 correspond to steps S302 to S310 in the above embodiment, and the five components have the same instances and application scenarios as the corresponding steps, but are not limited to the content disclosed in the above embodiment. It should be noted that the above components can be hardware components or software components stored in a memory (for example, the memory 2604) and processed by one or more processors (for example, the processors 2602a, 2602b, …, 2602n), or the above components can be a part of the apparatus and can run in the computer terminal A provided in the embodiment of the present disclosure. According to the embodiment of the present disclosure, an interactive information processing apparatus for implementing the interactive information processing method shown in FIG. 4 is further provided. FIG. 24 is a schematic diagram of another interactive information processing apparatus according to an embodiment of the present disclosure. As shown in FIG. 24, the interactive information processing apparatus 2400 can include a second determining component 2402 and a first creating component 2404.The second determining component 2402 is configured to determine an initial information generation rule required to be met by the information processing model in a process of generating reply information corresponding to the input information, where the initial information generation rule is used to guide the information processing model to determine a target information generation rule required to be met in the process of generating the reply information; and the first creating component 2404 is configured to create configuration information based on the initial information generation rule, where the configuration information is used to generate prompt information of the information processing model, and the prompt information is used to generate the reply information. It should be noted that the second determining component 2402 and the first creating component 2404 correspond to steps S402 to S404 in the foregoing embodiment, and the two components have the same instances and application scenarios as the corresponding steps, but are not limited to the content disclosed in the foregoing embodiment. It should be noted that the components can be hardware components or software components stored in a storage (for example, the memory 2604) and processed by one or more processors (for example, the processors 2602a, 2602b,..., 2602n), or the components can be a part of the apparatus and can run in the computer terminal A provided in the embodiment of the present disclosure. FIG. 25 is a schematic diagram of another interactive information processing apparatus according to the embodiment of the present disclosure. As shown in FIG. 25, the interactive information processing apparatus 2500 can include a fourth obtaining component 2502, a fifth obtaining component 2504, a fifth generating component 2506, and a sixth generating component 2508. The fourth obtaining component 2502 is configured to obtain query information, where the query information is used to represent information to be replied in a task-oriented dialogue scenario. The fifth obtaining component 2504 is configured to obtain configuration information associated with the query information, where the configuration information is used to describe an initial information generation rule required to be met by a dialogue model in a process of generating reply information corresponding to the query information. The initial information generation rule is used to guide the dialogue model to determine a target information generation rule required to be met in the process of generating the reply information, and the dialogue model is obtained by training based on query information samples in the task-oriented dialogue scenario and reply information samples in the task-oriented dialogue scenario. The fifth generating component 2506 is configured to generate prompt information of the dialogue model based on the configuration information and pre-obtained parameter information associated with the query information, where the parameter information is used to define key content in the reply information. The sixth generating component 2508 is configured to guide the dialogue model to generate the reply information including the key content according to the initial information generation rule and the target information generation rule by using the prompt information.The fourth obtaining component 2502, the fifth obtaining component 2504, the fifth generating component 2506, and the sixth generating component 2508 correspond to steps S502-S508 in the above-described embodiments, and the four components have the same instances and application scenarios as the corresponding steps, but are not limited to the above-described embodiments. It should be noted that the components can be hardware components or software components stored in a memory (for example, the memory 2604) and processed by one or more processors (for example, the processors 2602a, 2602b,..., 2602n), or the components can be run in the computer terminal A provided in the embodiments of the present disclosure as part of the device. In the processing device of the interactive information, if the input information to be replied to needs to be replied to, the input information can be obtained, and the input information can be input into the information processing model to process the input information. In the processing process, the initial information generation rule required to be met by the input information, that is, the configuration information of the information processing model, can be determined. The configuration information and the corresponding parameter information obtained in advance can be used to generate the prompt information of the reply information processing model to process the input information and generate the prompt information of the reply information. Through the prompt information, the information processing model determines the target information generation rule required to be met in the process of generating the reply information, so that the reply information containing the key content can be generated according to the target information generation rule. Since the operation personnel is required to participate in the configuration of the dialogue system, which can cause the problem of low efficiency of the entire interactive information processing, the embodiments of the present disclosure can define configuration information related to the task flow of the processing of the interactive information in the dialogue system, the configuration information is easy to master and maintain, and the configuration information can effectively avoid the interference information to the information processing model. Based on this, the embodiments of the present disclosure rely on the reasoning ability and instruction compliance of the information processing model, use the prompt information generated by the configuration information, so that the information processing model can determine the target information generation rule in the process of generating the reply information, that is, perform autonomous planning of the flow, and does not need to define the corresponding dialogue logic artificially, so that the information processing model can smoothly and naturally generate the reply information of the reply input information, thereby achieving the technical effect of improving the processing efficiency of the interactive information, and solving the technical problem of low processing efficiency of the interactive information. The embodiments of the present disclosure can provide a computer terminal, which can be any one of the computer terminal devices in the computer terminal group. Optionally, in the embodiments, the computer terminal can be replaced by a terminal device such as a mobile terminal. Optionally, in the embodiments, the computer terminal can be located in at least one network device of a plurality of network devices in a computer network.In the embodiment, the computer terminal can execute program codes of the following steps in the method for processing interactive information: obtaining input information to be replied; obtaining configuration information associated with the input information, wherein the configuration information is used to describe initial information generation rules required to be met in a process in which an information processing model generates reply information corresponding to the input information, the initial information generation rules are used to guide the information processing model to determine target information generation rules required to be met in the process of generating the reply information; generating prompt information of the information processing model based on the configuration information and parameter information associated with the input information and obtained in advance, wherein the parameter information is used to define key content in the reply information; and generating the reply information including the key content by using the prompt information. Optionally, FIG. 26 is a structural block diagram of a computer terminal according to an embodiment of the present disclosure. As shown in FIG. 26, the computer terminal A can include one or more (only one is shown in the figure) processors 2602, a memory 2604, and a transmission device 2606. The memory can be used to store software programs and components, such as program instructions / components corresponding to the method and device for processing interactive information in the embodiment of the present disclosure. The processor executes various function applications and data processing by running the software programs and components stored in the memory, that is, implements the method for processing interactive information. The memory can include a high-speed random access memory, and can further include a nonvolatile memory, such as one or more magnetic storage devices, a flash memory, or other nonvolatile solid-state memories. In some examples, the memory can further include a memory remotely arranged with respect to the processor, and the remote memory can be connected to the computer terminal A through a network. Examples of the network include but are not limited to the Internet, an enterprise intranet, a local area network, a mobile communication network, and a combination thereof. Optionally, the processor can further execute program codes of the following steps: performing vectorization processing on the input information to obtain a first vector; querying, from a vector database, a second vector having a similarity greater than a similarity threshold value with the first vector; and determining configuration information corresponding to the second vector. Optionally, the processor can further execute program codes of the following steps: determining a scenario to which the first vector belongs; querying, from a vector set corresponding to the scenario in the vector database, a second vector having a similarity greater than a similarity threshold value with the first vector, and / or at least one vector corresponding to a calling interface of the scenario.Optionally, the processor can further execute program codes of the following steps: if the second vector includes a vector corresponding to a flow of the scene, determining an initial execution flow of the scene as an initial execution flow required in a generation process of the reply information, to obtain the configuration information; if the second vector includes a vector corresponding to a calling interface of the scene, determining the calling interface of the scene as an initial calling interface required in the generation process of the reply information, to obtain the configuration information. Optionally, the processor can further execute program codes of the following steps: obtaining parameter information from the server; combining the configuration information and the parameter information to obtain the prompt information of the information processing model. Optionally, the processor can further execute program codes of the following steps: obtaining at least one of the following information as auxiliary information: task description information under a scene to which the input information belongs, environment information under the scene to which the input information belongs, context information of the input information; combining the configuration information, the parameter information and the auxiliary information to obtain the guide information. Optionally, the processor can further execute program codes of the following steps: controlling the information processing model to determine the intent data of the input information; generating the reply information according to the intent data and the prompt information. Optionally, the processor can further execute program codes of the following steps: controlling the information processing model to determine a parameter value of the parameter information in the prompt information according to the intent data; generating the reply information according to the parameter value of the parameter information and the prompt information. Optionally, the processor can further execute program codes of the following steps: in a case where the initial information generation rule includes an initial execution flow, controlling the information processing model to determine a parameter value of the parameter information matched with the initial execution flow according to the intent data. Optionally, the processor can further execute program codes of the following steps: controlling the information processing model to generate the reply information according to the initial information generation rule and the target information generation rule corresponding to the configuration information in the prompt information. Optionally, the processor can further execute program codes of the following steps: in a case where the initial information generation rule includes an initial execution flow, controlling the information processing model to execute according to the initial execution flow corresponding to the parameter value, to obtain a first execution result; controlling the information processing model to determine the target information generation rule according to the first execution result; controlling the information processing model to execute according to the target information generation rule, to obtain the reply information. Optionally, the processor can further execute program codes of the following steps: controlling the information processing model to determine instruction description information in the target execution flow according to the first execution result; and / or, controlling the information processing model to determine a target calling interface according to the first execution result.Optionally, the processor can further execute program codes of the following steps: in a case where the initial information generation rule comprises an initial calling interface, controlling the information processing model to call the initial calling interface for execution, to obtain a second execution result; controlling the information processing model to determine a target information generation rule by using the second execution result; and controlling the information processing model to execute according to the target information generation rule, to obtain the reply information. The processor can call information and application programs stored in the storage through the transmission device, to execute the following steps: obtaining input information to be replied; determining a scenario to which the input information belongs; obtaining configuration information associated with the input information in the scenario, wherein the configuration information is used to describe an initial information generation rule required to be met in a process of generating reply information corresponding to the input information, and the initial information generation rule is used to guide the information processing model to determine a target information generation rule required to be met in the process of generating the reply information; generating prompt information of the information processing model based on the configuration information and parameter information associated with the input information and obtained in advance; and generating the reply information that meets the scenario and comprises key content, by using the prompt information. The processor can call information and application programs stored in the storage through the transmission device, to execute the following steps: determining an initial information generation rule required to be met by the information processing model in a process of generating reply information corresponding to input information, wherein the initial information generation rule is used to guide the information processing model to determine a target information generation rule required to be met in the process of generating the reply information; and creating configuration information based on the initial information generation rule, wherein the configuration information is used to generate prompt information of the information processing model, and the prompt information is used to generate the reply information. The processor can call information and application programs stored in the storage through the transmission device, to execute the following steps: obtaining inquiry information, wherein the inquiry information is used to represent information to be replied in a task-oriented dialogue scenario; obtaining configuration information associated with the inquiry information, wherein the configuration information is used to describe an initial information generation rule required to be met in a process of generating reply information corresponding to the inquiry information, and the initial information generation rule is used to guide the dialogue model to determine a target information generation rule required to be met in the process of generating the reply information; generating prompt information of the dialogue model based on the configuration information and parameter information associated with the inquiry information and obtained in advance, wherein the parameter information is used to define key content in the reply information; and guiding the dialogue model to generate the reply information comprising the key content according to the initial information generation rule and the target information generation rule, by using the guide information. The present disclosure provides a processing method of interactive information.In the embodiment of the present disclosure, if the input information to be replied needs to be replied, the input information can be acquired, and the input information can be input into the information processing model to process the input information. In the processing process, the initial information generation rule required to be met by the input information, that is, the configuration information of the information processing model, can be determined. The prompt information for guiding the information processing model to process the input information to generate the reply information can be generated based on the configuration information and the corresponding parameter information acquired in advance. Through the prompt information, the target information generation rule required to be met in the process of generating the reply information by the information processing model is determined. Therefore, the reply information containing the key content can be generated according to the target information generation rule. Since the operation personnel is required to participate in the configuration of the dialogue system too much, which can cause the problem of reducing the efficiency of the whole interactive information processing, the embodiment of the present disclosure can define a configuration information related to the task flow of the interactive information processing in the dialogue system, the configuration information is easy to master and maintain, and the configuration information can effectively avoid giving a large amount of interference information to the information processing model. Based on this, the embodiment of the present disclosure relies on the reasoning ability and instruction compliance of the information processing model, and uses the prompt information generated by the configuration information, so that the information processing model can determine the target information generation rule in the process of generating the reply information, that is, the flow is autonomously planned, and the corresponding dialogue logic does not need to be defined artificially, so that the purpose that the information processing model can smoothly and naturally generate the reply information of the reply input information is achieved, and the technical effect of improving the processing efficiency of the interactive information is realized, and the technical problem of low processing efficiency of the interactive information is solved. As can be understood by those skilled in the art, the structure shown in FIG. 26 is only schematic, and the computer terminal A can also be a terminal device such as a smart phone (such as an Android phone, an iOS phone, etc.), a tablet computer, a palm computer, a mobile Internet device (Mobile Internet Devices, abbreviated as MID), a personal access display (Personal Access Display, abbreviated as PAD), etc. FIG. 26 does not limit the structure of the above computer terminal A. For example, the computer terminal A can further include more or fewer components (such as a network interface, a display device, etc.) than those shown in FIG. 26, or have a different configuration from that shown in FIG. 26.Those skilled in the art can understand that all or part of the steps of various methods in the above embodiments can be completed by instructing the hardware related to the terminal device by a program, and the program can be stored in a computer readable storage medium, and the storage medium can include a flash disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc. The embodiments of the present disclosure further provide a computer readable storage medium. Alternatively, in the present embodiment, the above computer readable storage medium can be used to save the program code executed by the processing method of the interactive information provided by the first embodiment. Alternatively, in the present embodiment, the above computer readable storage medium can be located in any one of the computer terminal group in the computer network or any one of the mobile terminal group. Alternatively, in the present embodiment, the computer readable storage medium is configured to store the program code for executing the following steps: obtaining input information to be replied; obtaining configuration information associated with the input information, wherein the configuration information is used to describe the initial information generation rule required to be met in the process of generating the reply information corresponding to the input information by the information processing model, and the initial information generation rule is used to guide the information processing model to determine the target information generation rule required to be met in the process of generating the reply information; generating prompt information of the information processing model based on the configuration information and the parameter information corresponding to the reply information, wherein the parameter information is used to define the key content in the reply information; and generating the reply information including the key content by using the prompt information. Alternatively, the above computer readable storage medium can further execute the program code of the following steps: performing vectorization processing on the input information to obtain a first vector; querying, from a vector database, a second vector having a similarity greater than a similarity threshold value with the first vector; and determining the configuration information corresponding to the second vector. Alternatively, the above computer readable storage medium can further execute the program code of the following steps: determining a scene to which the first vector belongs; and querying, in a vector set corresponding to the scene in the vector database, a second vector having a similarity greater than a similarity threshold value with the first vector, and / or at least one vector corresponding to the calling interface of the scene.Optionally, the computer readable storage medium can further execute program codes of the following steps: if the second vector includes a vector corresponding to a flow of the scenario, determining an execution flow of the scenario as an initial execution flow required in a generation process of the reply information, to obtain the configuration information; if the second vector includes a vector corresponding to a calling interface of the scenario, determining a calling interface of the scenario as an initial calling interface required in the generation process of the reply information, to obtain the configuration information. Optionally, the computer readable storage medium can further execute program codes of the following steps: obtaining parameter information from the server; and combining the configuration information and the parameter information to obtain the prompt information of the information processing model. Optionally, the computer readable storage medium can further execute program codes of the following steps: obtaining at least one of the following information as auxiliary information: task description information under a scenario to which the input information belongs, environment information under the scenario to which the input information belongs, and context information of the input information; and combining the configuration information, the parameter information, and the auxiliary information to obtain the guide information. Optionally, the computer readable storage medium can further execute program codes of the following steps: controlling the information processing model to determine the intent data of the input information; and generating the reply information according to the intent data and using the prompt information. Optionally, the computer readable storage medium can further execute program codes of the following steps: controlling the information processing model to determine a parameter value of the parameter information in the prompt information according to the intent data; and generating the reply information according to the parameter value of the parameter information and using the prompt information. Optionally, the computer readable storage medium can further execute program codes of the following steps: in a case where the initial information generation rule includes the initial execution flow, controlling the information processing model to determine a parameter value of the parameter information that matches the initial execution flow according to the intent data. Optionally, the computer readable storage medium can further execute program codes of the following steps: controlling the information processing model to generate the reply information according to the initial information generation rule and the target information generation rule corresponding to the configuration information in the prompt information. Optionally, the computer readable storage medium can further execute program codes of the following steps: in a case where the initial information generation rule includes the initial execution flow, controlling the information processing model to perform according to the initial execution flow corresponding to the parameter value to obtain a first execution result; controlling the information processing model to determine the target information generation rule using the first execution result; and controlling the information processing model to perform according to the target information generation rule to obtain the reply information. Optionally, the computer readable storage medium can further execute program codes of the following steps: controlling the information processing model to determine instruction description information in the target execution flow using the first execution result; and / or, controlling the information processing model to determine the target calling interface using the first execution result.Optionally, the computer readable storage medium can further execute program codes of the following steps: in a case that the initial information generation rule comprises an initial calling interface, controlling the information processing model to call the initial calling interface to execute, to obtain a second execution result; controlling the information processing model to determine a target information generation rule by using the second execution result; and controlling the information processing model to execute according to the target information generation rule to obtain the reply information. As an optional example, the computer readable storage medium is configured to store program codes for executing the following steps: obtaining input information to be replied; determining a scenario to which the input information belongs; obtaining configuration information associated with the input information in the scenario, wherein the configuration information is used to describe an initial information generation rule required to be met by the information processing model in a process of generating reply information corresponding to the input information, the initial information generation rule is used to guide the information processing model to determine a target information generation rule required to be met in the process of generating the reply information; generating prompt information of the information processing model based on the configuration information and parameter information associated with the input information and obtained in advance; and generating the reply information that meets the scenario and comprises key content by using the prompt information. As an optional example, the computer readable storage medium is configured to store program codes for executing the following steps: determining an initial information generation rule required to be met by the information processing model in a process of generating reply information corresponding to input information, wherein the initial information generation rule is used to guide the information processing model to determine a target information generation rule required to be met in the process of generating the reply information; and creating configuration information based on the initial information generation rule, wherein the configuration information is used to generate prompt information of the information processing model, and the prompt information is used to generate the reply information. As an optional example, the computer readable storage medium is configured to store program codes for executing the following steps: obtaining inquiry information, wherein the inquiry information is used to represent information to be replied in a task-oriented dialogue scenario; obtaining configuration information associated with the inquiry information, wherein the configuration information is used to describe an initial information generation rule required to be met by the dialogue model in a process of generating reply information corresponding to the inquiry information, the initial information generation rule is used to guide the dialogue model to determine a target information generation rule required to be met in the process of generating the reply information; generating prompt information of the dialogue model based on the configuration information and parameter information associated with the inquiry information and obtained in advance, wherein the parameter information is used to define key content in the reply information; and guiding the dialogue model to generate the reply information comprising the key content according to the initial information generation rule and the target information generation rule by using the prompt information. The embodiments of the present disclosure further provide a computer program product.Optionally, in the embodiment, the computer program product can include a computer program, and the computer program, when executed by a processor, implements the method provided by the above embodiment. Optionally, the computer program product can include a nonvolatile computer readable storage medium, and the nonvolatile computer readable storage medium can be used to store a computer program, and the computer program, when executed by a processor, implements the method provided by the above embodiment. Embodiments of the present disclosure can provide an electronic device, which can include a memory and a processor. FIG. 27 is a block diagram of an electronic device for processing interactive information according to an embodiment of the present disclosure. The electronic device is intended to represent various forms including digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device can also represent various forms of mobile devices, such as personal digital assistants, cellular telephones, smart phones, wearable devices, and other similar computing devices. The components shown in the figure, their connections and relationships, and their functions, are meant to be examples only, and are not meant to limit the described and / or claimed implementations of the present disclosure as set forth herein. As shown in FIG. 27, device 2700 includes computing assembly 2701, which can perform various appropriate actions and processes according to computer programs stored in read-only memory (ROM) 2702 or loaded into random access memory (RAM) 2703 from storage assembly 2708. In RAM 2703, various programs and data required for operation of device 2700 can also be stored. Computing assembly 2701, ROM 2702, and RAM 2703 are connected to each other via bus 2704. Input / output (I / O) interface 2705 is also connected to bus 2704. Various components in device 2700 are connected to I / O interface 2705, including: input assembly 2706, such as a keyboard, a mouse, etc.; output assembly 2704, such as various types of displays, a speaker, etc.; storage assembly 2708, such as a magnetic disk, an optical disk, etc.; and communication assembly 2709, such as a network card, a modem, a wireless communication transceiver, etc. Communication assembly 2709 allows device 2700 to exchange information / data with other devices via computer networks, such as the Internet, and / or various telecommunications networks. Computing assembly 2701 can be various general-purpose and / or special-purpose processing assemblies with processing and computing capabilities.Some examples of the computing component 2701 include, but are not limited to, a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), various dedicated Artificial Intelligence (Al) computing chips, various computing components running machine learning model algorithms, a Digital Signal Processor (DSP), and any suitable processor, controller, microcontroller, and so forth. The computing component 2701 performs various methods and processes described above, such as the data verification method. For example, in some embodiments, the data verification method can be implemented as a computer software program tangibly embodied in a machine-readable medium, such as the storage component 2708. In some embodiments, portions or all of the computer program can be loaded onto the device 2700 via the ROM 2702 and / or the communications component 2709. When the computer program is loaded onto the RAM 2703 and executed by the computing component 2701, one or more steps of the data verification method described above can be performed. Alternatively, in other embodiments, the computing component 2701 can be configured, by way of other any suitable approach (e.g., with the aid of firmware), to perform the data verification method. Various implementations of the systems and techniques described above herein can be realized in digital electronic circuitry, integrated circuitry, a Field Programmable Gate Array (FPGA), an Application Specific Integrated Circuit (ASIC), an Application Specific Standard Product (ASSP), a System-on-a-Chip (SOC), a Complex Programmable Logic Device (CPLD), computer hardware, firmware, software, and / or combinations thereof.The various embodiments described above may include: implementation in one or more computer programs, which can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device. Program code for implementing the methods of this disclosure may be written in any combination of one or more programming languages. The program code may be provided to a processor or controller of a general-purpose computer, a dedicated computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the functions / operations specified in the flowcharts and / or block diagrams are implemented. The program code may be executed entirely on a machine, partially on a machine, partially on a remote machine as a standalone software package, or entirely on a remote machine or server. The method embodiments provided in this disclosure can be executed on a mobile terminal, a computer terminal, or a similar computing device. Figure 28 is a hardware structure block diagram of a computer terminal (or mobile device) for implementing an interactive information processing method according to an embodiment of the present disclosure. As shown in Figure 28, the computer terminal 280 (or mobile device) may include one or more processors 2802 (shown as 2802a, 2802b, ..., 2802n in the figure) (processor 2802 may include, but is not limited to, a microprocessor (MCU) or a field programmable gate array (FPGA), etc.), a memory 2804 for storing data, and a transmission device 2806 for communication functions. In addition, it may also include: a display, an input / output interface (I / O interface), a Universal Serial Bus (USB) port (which may be included as one of the ports of a BUS bus), a network interface, a power supply, and / or a camera. Those skilled in the art will understand that the structure shown in FIG28 is merely illustrative and does not limit the structure of the electronic device described above. For example, the computer terminal 280 may also include more or fewer components than those shown in FIG28, or have a different configuration than that shown in FIG28.Figure 28 illustrates a hardware structure diagram, which can be used as an exemplary diagram of the above-mentioned computer terminal 280 (or mobile device) and as an exemplary diagram of the above-mentioned server. In an alternative embodiment, Figure 28 illustrates an embodiment in which the above-mentioned computer terminal 280 (or mobile device) shown in Figure 28 is used as a computing node in a computing environment 2901. Figure 29 is a structural diagram of a computing environment for a method of processing interactive information according to an embodiment of the present disclosure. As shown in Figure 29, the computing environment 2901 includes a plurality of computing nodes (e.g., servers) (represented by 2910-1, 2910-2, etc.) running on a distributed network. The computing nodes include local processing and memory resources, and an end user 2902 can remotely run applications or store data in the computing environment 2901. Applications can be provided as a plurality of services 2920-1, 2920-2, 2920-3, and 2920-4 in the computing environment 2901, representing services “I”, “D”, “E”, and “H”, respectively. The end user 2902 can provide and access the services through a web browser or other software applications on a client, and in some embodiments, the provision and / or requests of the end user 2902 can be provided to an entry gateway 2930. The entry gateway 2930 can include a corresponding proxy to handle the provision and / or requests for services (one or more services provided in the computing environment 2901). The services are provided or deployed according to various virtualization technologies supported by the computing environment 2901. In some embodiments, the services can be provided according to virtual machine (VM)-based virtualization, container-based virtualization, and / or the like. VM-based virtualization can simulate a real computer by initializing a virtual machine to execute programs and applications without directly accessing any actual hardware resources. While the virtual machine is virtualized, according to container-based virtualization, a container can be started to virtualize an entire operating system (OS) so that multiple workloads can run on a single operating system instance. In an embodiment of container-based virtualization, several containers of a service can be assembled into a Pod (e.g., a Kubernetes Pod). For example, as shown in Figure 29. Pod 2940-1, 2940-2, 2940-N (collectively, Pods). A Pod can include a proxy 2945 and one or more containers 2942-1, 2942-2,, 2942-M (collectively, containers). The one or more containers in a Pod handle requests related to one or more respective functions of a service, and the proxy 2945 generally controls network functions related to the service, such as routing, load balancing, etc. Other services can also be equipped with a Pod similar to the Pods. During operation, executing a user request from an end user 2902 can require invoking one or more services in the computing environment 2901, and executing one or more functions of a service can require invoking one or more functions of another service. As shown in FIG. 29, a service “I” 2920-1 receives a user request from an end user 2902 from an ingress gateway 2930, the service “I” 2920-1 can invoke a service “D” 2920-2, which can request a service “E” 2920-3 to perform one or more functions. The computing environment described above can be a cloud computing environment, where allocation of resources is managed by a cloud service provider, allowing development of functionality without considering implementation, tuning, or scaling servers. The computing environment allows developers to execute code that responds to events without building or maintaining complex infrastructure. Services can be split into a set of functions that can automatically scale independently, rather than scaling a single hardware device to handle potential loads. In the context of this disclosure, a machine-readable medium can be a tangible medium that can contain or store program code for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable storage media can include but are not limited to electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media can include an electrical connection based on one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having a display device (e.g., a Cathode Ray Tube (CRT) or a Liquid Crystal Display (LCD) monitor) for displaying information to the user and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the computer. Other kinds of devices can be used to provide for interaction with a user as well; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form, including acoustic, speech, or tactile input. The systems and techniques described here can be implemented in a computing system that includes a back end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front end component (e.g., a client computer having a graphical user interface or a Web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of the above. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network (LAN), a wide area network (WAN), and the Internet. The computing system can include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server can be a cloud server, a server of a distributed system, or a server combined with a blockchain. It should be noted that the above-mentioned sequence numbers of the embodiments of the present disclosure are only for description, and do not represent the advantages and disadvantages of the embodiments. In the above-mentioned embodiments of the present disclosure, the description of each embodiment is focused on, and the parts not described in detail in a certain embodiment can be referred to the related description of other embodiments. In the several embodiments provided by the present disclosure, it should be understood that the disclosed technical contents can be implemented by other means.Wherein, the apparatus embodiments described above are merely illustrative, for example, the division of components is only a logical function division, and actual implementation can have another division manner, for example, multiple components or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the components shown or discussed can be through some interfaces, indirect coupling or communication connection of components, and can be electrical or other forms. The components described as separate components can be or can not be physically separated, and the components shown as components can be or can not be physical components, that is, can be located in one place, or can be distributed on multiple network components. Part or all of the components can be selected to achieve the purpose of the embodiment scheme according to actual needs. In addition, the functional components in each embodiment of the disclosure can be integrated in a processing component, or each component can be physically present alone, or two or more components can be integrated in one component. The integrated component can be realized in the form of hardware or software functional component. When the integrated component is realized in the form of software functional component and sold or used as an independent product, it can be stored in a computer readable storage medium. Based on the above understanding, the technical scheme of the disclosure or the whole or part of the contribution to the prior art can be embodied in the form of a software product, which is stored in a storage medium and includes instructions to make a computer device (which can be a personal computer, a server or a network device, etc.) execute all or part of the steps of the method of each embodiment of the disclosure. The foregoing storage medium includes: U disk, read-only memory, random access memory, mobile hard disk, magnetic disk or optical disk and various program code storage media. The above is only the preferred embodiment of the disclosure, and it should be pointed out that for ordinary skilled persons in the art, without departing from the principles of the disclosure, some improvements and refinements can be made, and the above improvements and refinements are also regarded as the protection scope of the disclosure.The scheme provided by the present disclosure can be applied to the processing of interactive information, input information to be replied can be obtained, configuration information associated with the input information can be obtained, the configuration information is used to describe an initial information generation rule required to be met in a process in which an information processing model generates reply information corresponding to the input information, the initial information generation rule is used to guide the information processing model to determine a target information generation rule required to be met in the process of generating the reply information, the prompt information of the information processing model can be generated based on the configuration information and pre-acquired parameter information associated with the input information, the parameter information is used to define key content in the reply information, and the reply information including the key content can be generated by using the prompt information. By defining a configuration information related to a task flow of processing of interactive information in a dialogue system, the configuration information is easy to master and maintain, and a large amount of interference information can be effectively avoided from being given to the information processing model. Based on this, the present disclosure relies on the reasoning capability and instruction compliance of the information processing model, uses the prompt information generated by the configuration information, so that the information processing model can determine the target information generation rule in the process of generating the reply information, that is, the flow is autonomously planned, and the corresponding dialogue logic does not need to be defined by a person, thereby achieving the purpose that the information processing model can smoothly and naturally generate the reply information of the reply information of the input information, and further achieving the technical effect of improving the processing efficiency of the interactive information, and solving the technical problem of low processing efficiency of the interactive information.

Claims

Claims 1. A method for processing interactive information, comprising: Obtain the input information to be answered; obtain the configuration information associated with the input information, wherein the configuration information describes the initial information generation rules that the information processing model needs to satisfy in the process of generating the response information corresponding to the input information, and the initial information generation rules are used to guide the information processing model to determine the target information generation rules to be satisfied in the process of generating the response information; based on the configuration information and the pre-obtained parameter information associated with the input information, generate the prompt information of the information processing model, wherein the parameter information is used to define the key content in the response information; use the prompt information to generate the response information including the key content.

2. The method according to claim 1, wherein, The initial information generation rules include an initial execution flow and / or an initial call interface; the target information generation rules include a target execution flow and / or a target call interface.

3. The method according to claim 1, wherein, The similarity between the input information and the configuration information is greater than the similarity threshold.

4. The method according to claim 3, wherein, The step of obtaining the configuration information associated with the input information includes: performing vectorization processing on the input information to obtain a first vector; querying a second vector from a vector database whose similarity to the first vector is greater than the similarity threshold, wherein the vector database is used to store multiple vectors, and different vectors are used to represent the results of vectorization processing on different configuration information; and determining the configuration information corresponding to the second vector.

5. The method according to claim 4, wherein, The step of querying a second vector in the vector database whose similarity to the first vector is greater than the similarity threshold includes: determining the scene to which the first vector belongs; and querying a second vector in the vector database corresponding to the scene whose similarity to the first vector is greater than the similarity threshold, wherein the vector set includes at least one vector corresponding to a process belonging to the scene, and / or at least one vector corresponding to an interface belonging to the scene.

6. The method according to claim 5, wherein, Determining the configuration information corresponding to the second vector includes: in response to the second vector including a vector corresponding to a process belonging to the scenario, setting the execution of the scenario... 62 The execution flow is determined as the initial execution flow required to be executed during the generation of the response information, and the configuration information is obtained; in response to the second vector including the vector corresponding to the calling interface belonging to the scenario, the calling interface of the scenario is determined as the initial calling interface required to be called during the generation of the response information, and the configuration information is obtained.

7. The method according to claim 1, wherein, The step of generating prompt information for the information processing model based on the configuration information and pre-acquired parameter information associated with the input information includes: obtaining the parameter information from the server; and combining the configuration information and the parameter information to obtain the prompt information for the information processing model.

8. The method according to claim 7, wherein, The method further includes: obtaining at least one of the following information as auxiliary information: task description information of the scenario to which the input information belongs, environmental information of the scenario to which the input information belongs, and context information of the input information; and combining the configuration information and the parameter information to obtain the prompt information of the information processing model, including: combining the configuration information, the parameter information, and the auxiliary information to obtain the prompt information.

9. The method according to claim 1, wherein, The step of generating the response information including the key content using the prompt information includes: controlling the information processing model to determine the intent data of the input information, wherein the intent data is used to describe the purpose expressed by the input information; and generating the response information according to the intent data using the prompt information.

10. The method of claim 9, wherein, The step of generating the response information based on the intent data and the prompt information includes: controlling the information processing model to determine the parameter values ​​of the parameter information in the prompt information according to the intent data; and generating the response information based on the parameter values ​​of the parameter information and the prompt information.

11. The method according to claim 10, wherein, The step of controlling the information processing model to determine the parameter values ​​of the parameter information in the prompt information according to the intent data includes: responding to the initial information generation rule including an initial execution flow, controlling the information processing model to determine the parameter values ​​of the parameter information that match the initial execution flow according to the intent data.

12. The method according to claim 10, wherein, The step of generating the response information based on the parameter values ​​of the parameter information and the prompt information includes: controlling the information processing model to generate the response information according to the initial configuration information corresponding to the prompt information. 63 The information generation rules and the target information generation rules are used to generate the response information.

13. The method according to claim 12, wherein, The method of controlling the information processing model to generate the response information according to the initial information generation rule and the target information generation rule corresponding to the configuration information in the prompt information includes: responding to the initial information generation rule including an initial execution flow, controlling the information processing model to execute according to the initial execution flow corresponding to the parameter value to obtain a first execution result, wherein the initial information generation rule includes the initial execution flow; controlling the information processing model to determine the target information generation rule using the first execution result; and controlling the information processing model to execute according to the target information generation rule to obtain the response information.

14. The method according to claim 13, wherein, The step of controlling the information processing model to determine the target information generation rule using the first execution result includes: controlling the information processing model to determine instruction description information in the target execution flow using the first execution result, wherein the instruction description information is used to describe instructions that the information processing model is allowed to execute; and / or, controlling the information processing model to determine the target calling interface using the first execution result.

15. The method according to claim 12, wherein, The method of controlling the information processing model to obtain the response information according to the determined initial information generation rule and target information generation rule corresponding to the configuration information includes: responding to the initial information generation rule including an initial call interface, controlling the information processing model to call the initial call interface for execution to obtain a second execution result; controlling the information processing model to determine the target information generation rule using the second execution result; and controlling the information processing model to execute according to the target information generation rule to obtain the response information.

16. A method for processing interactive information, comprising: Obtain the input information to be answered; determine the scenario to which the input information belongs; The system acquires configuration information associated with the input information in the given scenario. This configuration information describes the initial information generation rules that the information processing model must satisfy when generating the response information corresponding to the input information. These initial information generation rules guide the information processing model in determining the target information generation rules to be satisfied during the generation of the response information. Based on the configuration information and pre-acquired parameter information associated with the input information, the system generates prompt information for the information processing model. The parameter information defines the key content in the response information. Using the prompt information, the system generates response information that satisfies the scenario and includes the key content.

17. A method for processing interactive information, comprising: 64 Determine the initial information generation rules that the information processing model must satisfy in the process of generating response information corresponding to input information, wherein, The initial information generation rule is used to guide the information processing model to determine the target information generation rule to be satisfied during the process of generating the response information; based on the initial information generation rule, configuration information is created, wherein the configuration information is used to generate prompt information for the information processing model, and the prompt information is used to generate the response information.

18. The method according to claim 17, wherein, The initial information generation rules include an initial execution flow and / or an initial call interface; the target information generation rules include a target execution flow and / or a target call interface.

19. The method according to claim 17, wherein, The method further includes: creating parameter information corresponding to the response information, wherein the parameter information is used to define the key content in the response information, and the configuration information and the parameter information are used to generate the prompt information of the information processing model.

20. The method according to claim 17, wherein, The method further includes: vectorizing the different configuration information to obtain different vectors; and storing the different vectors in a vector database.

21. A method for processing interactive information, comprising: The process involves: acquiring query information, where the query information represents information to be answered in a task-oriented dialogue scenario; acquiring configuration information associated with the query information, where the configuration information describes the initial information generation rules that the dialogue model needs to satisfy when generating the response information corresponding to the query information, the initial information generation rules guiding the dialogue model to determine the target information generation rules to be satisfied when generating the response information, the dialogue model being trained based on query information samples and response information samples in the task-oriented dialogue scenario; generating prompt information for the dialogue model based on the configuration information and pre-acquired parameter information associated with the query information, where the parameter information defines the key content in the response information; and using the prompt information to guide the dialogue model to generate the response information including the key content according to the initial information generation rules and the target information generation rules.

22. The method of claim 21, wherein, The inquiry information and the response information are multimodal information, and the types of the multimodal information include at least one of the following: text information containing character information, video frame information containing frame image information, and audio information. The types of the response information include at least one of the following: text information, image information, video information, and voice information.

23. A dialogue system, comprising: The client is used to obtain input information that needs to be answered; The server is used to send configuration information associated with the input information to the client, wherein... 65 The configuration information describes the initial information generation rules that the information processing model needs to meet during the process of generating response information corresponding to the input information. These initial information generation rules guide the information processing model to determine the target information generation rules to be met during the generation of the response information. The client is used to generate prompt information for the information processing model based on the configuration information and pre-acquired parameter information associated with the input information, and uses the prompt information to generate response information including key content. The parameter information is used to define the key content in the response information.

24. The dialogue system according to claim 23, wherein, The server includes: an information configuration device, configured to vectorize different configuration information to obtain vectors of different configuration information, and store the vectors of different configuration information in a vector database; a client, configured to vectorize the input information to obtain a first vector, and send an information query request to the information configuration device, wherein the information query request includes the first vector; and the information configuration device, in response to the information query request, querying from the vector database a second vector whose similarity to the first vector is greater than a similarity threshold, and sending the configuration information corresponding to the second vector to the client.

25. An electronic device, comprising: Memory, which stores executable programs; A processor for running the program, wherein the program, when running, performs the method according to any one of claims 1 to 22.

26. A computer-readable storage medium comprising a stored executable program, wherein, when the executable program is executed, it controls the device on which the computer-readable storage medium is located to perform the method of any one of claims 1 to 22.

27. A computer program product comprising a computer program that, when executed by a processor, implements the method of any one of claims 1 to 22.

28. A computer program product, comprising: A computer program that, when executed by a processor, implements the method described in any one of claims 1 to 22.

29. A computer program product, comprising: A non-volatile computer-readable storage medium storing a computer program that, when executed by a processor, implements the method according to any one of claims 1 to 22.

30. A computer program that, when executed by a processor, implements the method of any one of claims 1 to 22. 66

Images