Service auxiliary information expression method, device and equipment

Abstract

The embodiment of the invention discloses a business auxiliary information expression method, device and equipment. The scheme comprises the following steps: performing intention analysis on a providing node of the service auxiliary information to obtain a first intention analysis result, and analyzing each historical expression corresponding to the service auxiliary information according to the first intention analysis result to obtain a historical expression analysis result; performing intention analysis on the sensing node of the service auxiliary information to obtain a second intention analysis result, and determining a corresponding intention scene based on the second intention analysis result; for the intention scene, performing data optimization and / or data derivation on the historical expression according to the historical expression analysis result to obtain corresponding candidate expressions; and evaluating the candidate expressions, and selecting a corresponding target expression according to an evaluation result to express the service auxiliary information to the sensing node.

Description

TECHNICAL FIELD

[0001] The present specification relates to the technical field of Internet, and particularly relates to a business auxiliary information expression method, device and equipment. BACKGROUND

[0002] With the development of computer and Internet technology, many businesses can be executed online. In order to improve user experience, relevant business auxiliary information can be provided during business execution to provide corresponding help for users.

[0003] In the traditional scheme, the corresponding business auxiliary information is usually generated by manual combination of actual business scenarios, and the business auxiliary information is expressed to the user through a preset expression method. For example, the R&D personnel need to generate relevant scripts for each business scenario, and need to ensure that the scripts do not violate relevant regulations. At the same time, the corresponding display page is generated, and the corresponding elements are added in the page and the format of the script is adjusted in the display page according to the type of the business scenario, so as to generate the corresponding expression method.

[0004] Therefore, the business auxiliary information needs to have higher timeliness and reusability, and can be personalized expressed to the access user. SUMMARY

[0005] One or more embodiments of the present specification provide a business auxiliary information expression method, device, equipment and storage medium, to solve the technical problem that the business auxiliary information needs to have higher timeliness and reusability, and can be personalized expressed to the access user.

[0006] To solve the above technical problem, one or more embodiments of the present specification are implemented as follows: One or more embodiments of the present specification provide a business auxiliary information expression method, comprising: Performing intent analysis on a providing node of business auxiliary information to obtain a first intent analysis result, and analyzing each historical expression corresponding to the business auxiliary information according to the first intent analysis result to obtain a historical expression analysis result; Performing intent analysis on a perception node of business auxiliary information to obtain a second intent analysis result, and determining a corresponding intent scene based on the second intent analysis result; According to the historical expression analysis result, data optimization and / or data derivation are performed on the historical expression to obtain a corresponding candidate expression for the intent scene; The candidate expression is evaluated, and a target expression is selected according to the evaluation result, which is used to express the business auxiliary information to the perception node.

[0007] The one or more embodiments of the specification provide a service auxiliary information expression device, comprising: A service auxiliary information expression device, comprising: A first intention analysis module, which performs intention analysis on a providing node of service auxiliary information to obtain a first intention analysis result, and analyzes each historical expression corresponding to the service auxiliary information according to the first intention analysis result to obtain a historical expression analysis result; A second intention analysis module, which performs intention analysis on a perception node of service auxiliary information to obtain a second intention analysis result, and determines a corresponding intention scene based on the second intention analysis result; A data optimization and derivation module, which performs data optimization and / or data derivation on the historical expression according to the historical expression analysis result for the intention scene to obtain a corresponding candidate expression; An evaluation module, which evaluates the candidate expression and selects a corresponding target expression according to an evaluation result, which is used to express the service auxiliary information to the perception node.

[0008] The one or more embodiments of the specification provide a service auxiliary information expression device, comprising: At least one processor; and, A memory in communication connection with the at least one processor; wherein, The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to: Perform intention analysis on a providing node of service auxiliary information to obtain a first intention analysis result, and analyze each historical expression corresponding to the service auxiliary information according to the first intention analysis result to obtain a historical expression analysis result; Perform intention analysis on a perception node of service auxiliary information to obtain a second intention analysis result, and determine a corresponding intention scene based on the second intention analysis result; Perform data optimization and / or data derivation on the historical expression according to the historical expression analysis result for the intention scene to obtain a corresponding candidate expression; Evaluate the candidate expression and select a corresponding target expression according to an evaluation result, which is used to express the service auxiliary information to the perception node.

[0009] The one or more embodiments of the specification provide a non-volatile computer storage medium, which stores computer executable instructions, and the computer executable instructions are set to: Intent analysis is performed on the nodes that provide business auxiliary information to obtain a first intent analysis result. Based on the first intent analysis result, the historical expressions corresponding to the business auxiliary information are analyzed to obtain a historical expression analysis result. Intent analysis is performed on the perception nodes of business auxiliary information to obtain a second intent analysis result, and the corresponding intent scenario is determined based on the second intent analysis result; For the stated intent scenario, based on the historical expression analysis results, the historical expressions are optimized and / or derived to obtain corresponding candidate expressions; The candidate expressions are evaluated, and the corresponding target expression is selected based on the evaluation results to express the business assistance information to the perception node.

[0010] The above-described at least one technical solution adopted in one or more embodiments of this specification can achieve the following beneficial effects: By enabling intelligent agents to autonomously perform complex operations such as intent analysis, expression generation, and evaluation, manual intervention is reduced, lowering labor costs. The entire process is essentially automated, requiring minimal human involvement. Intent analysis, automated data optimization, and data derivation rapidly generate candidate expressions, responding more efficiently to business scenario needs and improving timeliness. Historical expression analysis results and relevant experience can be accumulated and reused, avoiding repetitive work and supporting subsequent expression generation and optimization, thus enhancing reusability. Based on precise matching of intent scenarios at perception nodes, expressions tailored to the needs of perception nodes are generated, improving the reach of business support information and achieving personalized expression. Target expressions are evaluated and selected to avoid inefficient expressions consuming resources, improving the overall quality of business support information and optimizing resource utilization. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in the embodiments or prior art of this specification, the drawings used in the description of the embodiments or prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 A flowchart illustrating a method for expressing business auxiliary information provided in one or more embodiments of this specification; Figure 2 A schematic diagram of the system to which a business auxiliary information expression method in an application scenario belongs, provided for one or more embodiments of this specification; Figure 3 This is a flowchart illustrating the expression of business auxiliary information in one or more embodiments of this specification. Figure 4 This is a flowchart illustrating a timed trigger scenario analysis task in an application scenario, provided for one or more embodiments of this specification. Figure 5 A schematic diagram of the structure of a business auxiliary information expression device provided in one or more embodiments of this specification; Figure 6 This is a schematic diagram of the structure of a business auxiliary information expression device provided for one or more embodiments of this specification. Detailed Implementation

[0013] This specification provides methods, apparatus, devices, and storage media for expressing business auxiliary information.

[0014] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this application.

[0015] Figure 1 This is a flowchart illustrating a method for expressing business auxiliary information provided in one or more embodiments of this specification. This method can be applied to different business domains, such as e-commerce, internet finance, instant messaging, gaming, government services, and operations. The process can be executed by computing devices in the corresponding domain (e.g., intelligent customer service servers or intelligent mobile terminals for payment services). Certain input parameters or intermediate results in the process can be manually adjusted to help improve accuracy.

[0016] Figure 1 The process may include the following steps: S102: Perform intent analysis on the node providing business auxiliary information to obtain a first intent analysis result, and analyze each historical expression corresponding to the business auxiliary information based on the first intent analysis result to obtain a historical expression analysis result.

[0017] Business support information refers to relevant information that assists users during the execution of business processes. It can include various types, such as window navigation components that guide users to the business execution page before execution; explanatory information for specific steps during execution; and navigation window components that guide users to related business processes and post-execution notes after completion. Business support information may also be referred to as supply, service, resource, or other descriptions in some contexts, depending on the specific business scenario.

[0018] Business support information also includes different content in different industries. For example, for the e-commerce industry, it may include redirection components for product detail pages, coupon redemption redirection components, help information on payment pages, and reminders about after-sales policies for products; for the financial industry, it may include risk warning information for wealth management products, step-by-step instructions for transaction processes, and notification components for changes in account balances.

[0019] The expression of business support information can also include multiple dimensions. For example, the expression location can include: a separately set dialog box, inside or above the current business execution page, or other pages of the application to which the current business belongs (including the application's splash screen, the application's homepage, and execution pages of other businesses, etc.). The elements included in the expression can also be presented in the form of text, images, audio, video, or a combination of modalities, depending on the requirements.

[0020] The term "providing node" refers to the node that formulates and generates the auxiliary information for the business and its corresponding expression. These nodes are usually developed, maintained, or operated by business-related or application-related R&D personnel, maintenance personnel, and operations personnel.

[0021] Intent analysis refers to the process of parsing and standardizing the intended actions that nodes want to perform by providing the content they express. Generally speaking, the results of the first intent analysis mainly include the goals that are hoped to be achieved through business auxiliary information, the core content to be conveyed, and the potential expectations for the expression method.

[0022] To achieve autonomy in the process of expressing business auxiliary information, multiple agents are pre-configured in the system to perform different functions. An agent is an entity (usually a software system) capable of autonomously perceiving the environment, making decisions, and achieving goals. It requires no human intervention throughout the process and can proactively adjust its behavior based on goals, environmental information, and historical experience to complete tasks.

[0023] A typical intelligent agent includes a perception module, a decision-making module, an action module, and a memory module. The perception module is used to acquire and parse environmental and user information, and to transform fuzzy input into structured information. The decision-making module is used to plan goals, break down tasks, and select action strategies. The action module is used to execute decisions. The memory module is used to store historical data, experience, or contextual information.

[0024] The intelligent agent incorporates corresponding technical tools to support each module in achieving its functions. These tools may include Large Language Models (LLMs), Neural Network Models (NNMs), pre-defined rules, and tool APIs. Figure 2 This diagram illustrates the framework of a system for a business auxiliary information expression method in an application scenario, provided in one or more embodiments of this specification. It establishes a main intelligent agent and multiple sub-intelligent agents. The main intelligent agent receives analysis results and feedback information from each sub-intelligent agent and, based on overall business objectives and current task progress, decomposes the business auxiliary information expression task into several sub-tasks, assigning them to corresponding sub-intelligent agents (e.g., intent analysis sub-intelligent agent, expression analysis sub-intelligent agent, expression optimization sub-intelligent agent, etc.). The responsibilities, execution order, and collaboration methods of each sub-intelligent agent are clearly defined. Under the unified scheduling of the main intelligent agent, each sub-intelligent agent focuses on completing its assigned specific task. For example, the intent analysis sub-intelligent agent is responsible for deeply analyzing the intents of providing nodes and sensing nodes, while the expression analysis sub-intelligent agent is responsible for multi-dimensional evaluation of historical expressions.

[0025] Figure 3 This specification provides a flowchart illustrating the expression of business auxiliary information in one or more embodiments, illustrating a process for expressing business auxiliary information in an application scenario. For the node providing the business auxiliary information, an intent analysis sub-agent performs intent analysis to obtain a first intent analysis result. The intent analysis sub-agent is primarily used to analyze the intent of the providing node, clarifying the providing node's opinions on the current expression of business auxiliary information, including evaluations of the current expression (including positive and negative evaluations) and the direction of adjustments to subsequent expressions. For example, if the providing node aims to improve the user appeal of the window jump component before business execution, and its original content is "I want to improve the user appeal of the window jump component in section B of channel A, thereby increasing the click-through rate of section B," an exemplary description of the first intent analysis result obtained through intent understanding could be: Channel: Channel A, Business Section: Section B, Operation Object: Business auxiliary information expression of the window jump component, Operation: Analysis, Generation, Optimization, Goal: Increase click-through rate.

[0026] The master agent orchestrates tasks based on the initial intent analysis results, resulting in the corresponding execution tasks for each sub-agent. Once the master agent understands the intent of the providing node, it can orchestrate tasks accordingly, specifying the tasks each sub-agent needs to perform.

[0027] Based on the execution task, historical expressions corresponding to business support information are obtained. Then, through an expression analysis sub-agent, these historical expressions are classified and analyzed according to corresponding business metrics to obtain historical expression analysis results. Historical expressions refer to the representation of business support information in history, which can be obtained through contextual information in historical records. The expression analysis sub-agent can retrieve historical expressions by accessing a database using appropriate tools or through online searches, and then perform data analysis using a large language model.

[0028] When acquiring historical expressions, the expression analysis sub-agent can acquire only historical expressions relevant to the current business based on the first intent analysis result. For example, for "I want to increase the user appeal of the window jump component in section B of channel A, thereby increasing the click-through rate of section B", it can acquire only the historical expressions of channel A or section B, or only the historical expressions of the window jump components of each section, or only the historical expressions related to improving user appeal, or combine the above historical expressions. It depends on the setting of the providing node and the judgment of the expression analysis sub-agent.

[0029] Business metrics are pre-set and primarily use quantitative methods to describe whether the current business information meets the corresponding standards. For example, for user attractiveness, the corresponding business metric could be click-through rate; for user experience, the corresponding business metrics could be user dwell time, business completion rate, user feedback, etc.

[0030] The expression analysis sub-agent classifies historical expressions based on business metrics. The classification can be divided into multiple categories according to whether the business metrics are met. Further refinement can be made within each category based on requirements. For example, in the category of not meeting business metrics, subcategories can be made such as some metrics not meeting business metrics, all metrics not meeting business metrics, and some metrics severely not meeting business metrics.

[0031] Of course, in addition to classification, the expression analysis sub-agent can also output corresponding analysis reports based on historical expression analysis results and store them in context. For example, it can store them in a Retrieval-Augmented Generation Knowledge Base (RAG).

[0032] Furthermore, when performing intent analysis through the intent analysis sub-agent, for the node providing business auxiliary information, intent analysis is performed through the general intent analysis sub-agent based on the interaction content with the providing node to obtain the general intent analysis result. This interaction content can be actively input by the providing node; for example, the providing node can input corresponding text content, which can be directly used as the interaction content. Alternatively, the intent analysis sub-agent can interact with the providing node through this text content to refine the text content and obtain the corresponding interaction content. Of course, the text content can be input by the providing node via text input, or it can be obtained by modal conversion after the providing node inputs audio or other modalities.

[0033] To ensure the accuracy of intent recognition, two sub-agents for intent analysis are set up: a general intent analysis sub-agent and a business auxiliary information-specific intent analysis sub-agent. Of course, corresponding intent analysis sub-agents can also be set up for other domains besides business auxiliary information expression, which will not be elaborated here.

[0034] The general intent analysis sub-agent is used to perform intent analysis on content across all industries and domains. It can distinguish between business auxiliary information analysis needs and business auxiliary information generation needs, and can perform scenario-based traffic splitting for the relevant business scenarios, as well as provide context-related follow-up questions to obtain relevant detailed content. This facilitates subsequent intent analysis by the dedicated intent analysis sub-agent for business auxiliary information. For example, if the interaction content is "I want to improve the user appeal of the window jump component in section B of channel A, thereby increasing the click-through rate of section B", the general intent analysis result identified by the general intent analysis sub-agent can be: primary intent: problem solving / troubleshooting, secondary intent: data analysis / processing, target object: section B, target object type: internal channel functional module, operation: improvement, attribute: click-through rate, constraint: needs to be optimized for specific scenarios of channel A, source: channel A.

[0035] At this point, although the intent of providing the node can be basically identified, some details are still not clear or accurate enough.

[0036] Based on this, and according to the general intent analysis results, intent analysis is performed through a dedicated intent analysis sub-agent for business auxiliary information to obtain the first intent analysis result. This dedicated intent analysis sub-agent can generate more professional intent analysis results by combining the detailed content of follow-up questions with the general intent analysis results. For example, if the interaction content is still "I want to improve the user appeal of the window jump component in section B of channel A, thereby increasing the click-through rate of section B," the final identified first intent analysis result could be: Main intent: metric improvement; Business auxiliary information object: expression of business auxiliary information; Channel: channel A; Business section: section B; Operation: analysis, generation, optimization. This makes the obtained first intent analysis result more concise, clear, and closer to the current scenario.

[0037] Figure 4 This document provides a flowchart illustrating a timed-triggered scenario analysis task in one or more embodiments of an application scenario. When generating interactive content, a scenario analysis task corresponding to the business auxiliary information expression scenario is executed based on a timed triggering method. For the business auxiliary information, to ensure its timely updates, it needs to be dynamically updated. Therefore, a corresponding time interval is set based on requirements to trigger the scenario analysis task. The duration of the time interval can be determined based on the data volume, current trending topics, etc.

[0038] Based on the scenario analysis task, sub-scenarios to be analyzed are obtained from various business auxiliary information expression scenarios. This determination process can be directly specified by the providing node; for example, only business auxiliary information expression scenarios in a specific industry or sector can be analyzed, and these can be used as sub-scenarios to be analyzed. Alternatively, the corresponding intelligent agent can automatically determine the sub-scenarios to be analyzed by combining the current popularity of each sub-scenarios (determined by click volume, attention, etc.) and the interval between the last analysis.

[0039] For each sub-expression scenario, a scenario analysis sub-agent determines its corresponding analysis dimensions and acquires the corresponding scenario data based on these dimensions for scenario analysis. The scenario analysis sub-agent analyzes the scenario data for each sub-expression scenario according to the analysis dimensions. These analysis dimensions can include multiple factors; for example, in the time dimension, analysis can be performed on sub-expression scenarios over 1 day, 7 days, and 30 days respectively. In the business dimension, scenario data is statistically analyzed based on different business needs. These business needs can include click-through rate, dwell time, and feedback satisfaction.

[0040] Based on the business requirements corresponding to each sub-expression scenario, the relevant scenario analysis results are filtered and assembled to generate interactive content that provides node output. Generally speaking, only the scenario analysis results that do not meet the business requirements can be selected for assembly to piece together the interactive content that provides node output.

[0041] S104: Perform intent analysis on the perception nodes of business auxiliary information to obtain a second intent analysis result, and determine the corresponding intent scenario based on the second intent analysis result.

[0042] A perception node refers to a node that can perceive the specific content expressed by business auxiliary information during business execution. It is typically the actual user performing the business, such as a consumer in a payment transaction or a buyer in an e-commerce transaction. Similar to performing intent analysis on the providing node, a corresponding intent analysis sub-agent can also be set up to perform intent analysis.

[0043] Specifically, such as Figure 2 and Figure 3 As shown, for the perception node providing business auxiliary information, intent analysis is performed by the intent analysis sub-agent to obtain a second intent analysis result. Unlike when intent analysis is performed on the providing node, due to the uncertainty of the perception node's intent, the second intent analysis result can be obtained by performing intent analysis solely through the general intent analysis sub-agent. Alternatively, a corresponding dedicated intent analysis sub-agent can be automatically selected based on the general intent analysis sub-agent to obtain the second intent analysis result.

[0044] At this point, the input to the intention analysis sub-agent may include the actions performed by the perception node, the input content, and historical action records.

[0045] The main agent orchestrates tasks based on the results of the second intent analysis, obtaining the execution tasks corresponding to each sub-agent. Similar to providing nodes, the main agent orchestrates the execution tasks.

[0046] Based on the task execution, the intent scenario of the perception node is obtained according to the second intent analysis result. Here, the intent scenario refers to the scenario corresponding to the second intent analysis result, which represents the intent of the perception node. Unlike the intent of the node providing the intent, its intent is usually not to improve the expression of business auxiliary information, but to seek the help of business auxiliary information, or to further enhance the user experience based on the current business by using business auxiliary information.

[0047] S106: For the intent scenario, based on the historical expression analysis results, perform data optimization and / or data derivation on the historical expressions to obtain corresponding candidate expressions.

[0048] Data optimization refers to improving the description, format, and structure of existing historical representations to provide a better business experience for sensing nodes.

[0049] Specifically, such as Figure 2 and Figure 3 As shown, for intent scenarios, based on the historical expression analysis results, the first original expression that does not meet business requirements is selected from the historical expressions. The business metrics corresponding to the first original expression do not meet business requirements and are unlikely to bring a high-quality business experience to users; therefore, it can also be called a poor-quality expression. If the business requirements have multiple dimensions, then it can be determined at what standard the number or degree of dimensions that fail to meet the business requirements must reach to be identified as the first original expression.

[0050] The expression optimization sub-agent optimizes the data of the first original expression to obtain the corresponding first candidate expression. This expression optimization sub-agent optimizes the data of the first original expression, including but not limited to description optimization, format optimization, and structure optimization. Its optimization targets can include text content, image content, audio content, and video content. For example, if the original transaction process description has poor user feedback and is considered a poor-quality first original expression, the expression optimization sub-agent updates the unclear parts of the original description to clearer descriptions, bolds some important content, and increases the contrast between the text and the background.

[0051] Unlike data optimization, data derivation does not directly optimize the original expression, but rather generates a new expression by understanding and analyzing the internal features of the original expression.

[0052] At this point, for the intent scenario, the expression generation sub-agent generates the corresponding second original expression. The expression generation sub-agent can directly generate new expressions; it can learn the internal features of existing high-quality original expressions through neural network models, large language models, etc., thereby generating new second original expressions.

[0053] The newly generated second original expression is difficult to match with the perception node because it is only generated by the expression generation sub-agent and does not involve the current intention scenario, or it can only have a low understanding of the current intention scenario.

[0054] Therefore, by using the crowd understanding sub-agent, a crowd profile matching the perception node is obtained based on the second intent analysis results. The crowd understanding sub-agent can obtain the corresponding user characteristics of the perception node through its registration information, historical behavior records, etc., and analyze the data of a specified crowd to produce general characteristics such as crowd preferences and interests. Based on these user characteristics, a corresponding user profile is determined, and the crowd is then categorized (e.g., through clustering algorithms). Based on the understanding, a list of crowds matching the consultation's preferences and interests is returned.

[0055] The expression selection sub-agent filters third original expressions from the second original expressions based on user profiles to obtain expressions that match the perception nodes. Different user profiles have different preferences for different expressions. The expression selection sub-agent can identify the expressions preferred by each user profile through large language models, neural network models, and preset matching rules, thus selecting more suitable third original expressions. The number and matching degree of the third original expressions can be set according to requirements. By using the expression selection sub-agent, high-quality third original expressions can be selected from a massive amount of second original expressions, thereby saving time in subsequent data derivation.

[0056] The expression generation sub-agent derives data from the third original expression to obtain the corresponding second candidate expression. The expression generation sub-agent extracts features from the third original expression, including text features, image features, scene constraint features, etc., which need to be determined based on the elements contained in the third original expression.

[0057] Since the third original expression itself matches the current intention scenario and has acquired high-quality experience from the second original expression, and also matches the crowd profile of the perception node, the expression generation sub-agent can generate a new second candidate expression that fits both the intention scenario and the crowd profile by understanding the logical features.

[0058] At this point, both the first and second candidate expressions obtained can be used as candidate expressions to be tested.

[0059] Furthermore, in the above scheme, the candidate expressions are generated after the sensing node accesses the business scenario. However, for some traffic fluctuation events, when the traffic of the sensing node is large, if the corresponding candidate expressions are still generated in real time, it may put a lot of pressure on the server.

[0060] Based on this, for the business scenarios corresponding to the business auxiliary information, traffic fluctuation events are identified according to the corresponding business indicators. A traffic fluctuation event refers to a fluctuation in the access traffic of a certain business scenario that has occurred or is about to occur exceeding a certain level. This can be monitored or predicted in real time based on the access volume of the business scenario or trending events on social media platforms.

[0061] In response to traffic fluctuation events, it is anticipated that a large number of sensing nodes may be about to access the business scenario, thus increasing the demand for business auxiliary information. However, not all groups will be interested in every traffic fluctuation event. Therefore, an event-based audience matching sub-agent is used to identify the audience profiles of designated groups whose match degree with the traffic fluctuation event is higher than a first preset level. The event-based audience matching sub-agent can calculate the match degree between the generated audience profiles and the traffic fluctuation event. The higher the match degree, the greater the probability that the group will access the business scenario. For example, in an e-commerce scenario, a traffic fluctuation event might be caused by the sudden surge in popularity of a certain sports product, leading to significant fluctuations in traffic to the product details page. In this case, through matching, it is believed that people who love exercise and are around 30-40 years old have a high match degree.

[0062] For each user profile, a third original expression is derived from the second original expression, and specified candidate expressions are derived from the data. Based on changes in business metrics during traffic fluctuation events, the time limit for the specified candidate expressions is dynamically set. At this point, instead of waiting for the user profile's corresponding business scenario to be accessed before generating candidate expressions, the corresponding candidate expressions are directly generated in advance, referred to as specified candidate expressions, and stored.

[0063] At the same time, a time limit can be set. Initially, the time limit can be set to a preset fixed value. As business indicators (such as access traffic, feedback evaluation indicators, etc.) remain high during traffic fluctuation events, the time limit can be appropriately extended until the business indicators decline to a certain level, at which point the time limit can no longer be extended.

[0064] Within the timeframe, if the current period is considered to be one of high traffic, and the perception node is identified as belonging to a specific user group based on its user profile, then the specified candidate expression is directly used as the second candidate expression. In this case, there's no need to generate a second candidate expression for the perception node through data derivation; instead, the already stored specified candidate expression is used directly, thus reducing server load.

[0065] Of course, for the first candidate expression, since it puts less pressure on the server's computing resources, it can be generated in advance or in real time as needed, and no restrictions are imposed here.

[0066] S108: Evaluate the candidate expressions and select the corresponding target expression based on the evaluation results to express the business assistance information to the perception node.

[0067] Specifically, such as Figure 2 and Figure 3As shown, the evaluation sub-agent performs internal evaluation of candidate expressions. The evaluation sub-agent can assess whether the generated candidate expressions conform to the scene specifications and whether the expression is clear, and it can also evaluate the analysis report generated above, such as which expressions have good guiding words, which grammatical guiding words are good, which guiding words are poor, and which grammatical guiding words are poor.

[0068] Since the evaluation process for the sub-agents has not yet been officially launched into business scenarios, it can be called internal evaluation.

[0069] For candidate expressions that pass internal evaluation, the configuration import sub-agent imports them into the business scenario corresponding to the business auxiliary information. The configuration import sub-agent can push the candidate expressions that pass internal evaluation to the relevant platform, where the providing node conducts manual review, and after approval, they are deployed to the business scenario.

[0070] Based on business scenarios, external evaluation is conducted on the imported candidate expressions. To ensure the stability of the external evaluation, A / B testing, canary testing, and other methods can be used. External evaluation is only conducted in a few business scenarios, and the candidate expressions that pass the external evaluation are used as the target expressions and deployed to all business scenarios on the platform.

[0071] Furthermore, if the business auxiliary information is used to jump to its corresponding business scenario, such as a business jump component before business execution, the frequent and identical expression of business auxiliary information may cause the perception node to feel fatigued, thereby reducing its interest in jumping to the business scenario.

[0072] Based on this, according to the evaluation results, several designated expressions with the highest business metrics are selected from each candidate expression, and these designated expressions are then sorted according to the business metrics. The evaluation results include the business metrics for each candidate expression. These business metrics are set according to the business requirements corresponding to the business scenario, and their contents have been described with examples above and will not be repeated here.

[0073] Acquire the perceived expressions of the sensing node within the most recent preset time period for the business scenarios corresponding to the business auxiliary information. The perceived expressions refer to the expressions of business auxiliary information that have been delivered to the sensing node, such as acquiring its perceived expressions within the most recent day.

[0074] For a given perceived expression, its similarity to each specified expression in the ranking representation is determined, and specified expressions with a similarity higher than a second preset level are filtered out. The similarity level can be determined by a weighted sum of similarities across multiple dimensions; for example, text similarity has the highest weight, image similarity has the second highest weight, and audio similarity has the lowest weight. This multi-dimensional comparison helps identify any specified expressions that are excessively similar to previously perceived expressions and filters them out, thereby increasing the novelty of the perceived node and enhancing its appeal for switching.

[0075] For the remaining specified expressions, select the corresponding specified expression as the target expression based on the ranking representation, which is used to convey business auxiliary information to the perception nodes. In this case, the specified expression with the highest ranking among the remaining specified expressions can be selected. Of course, if filtering results in the remaining specified expressions having poor business metrics, the specified expression with a higher degree of similarity can still be selected as the target expression.

[0076] By enabling intelligent agents to autonomously perform complex operations such as intent analysis, expression generation, and evaluation, manual intervention is reduced, lowering labor costs. The entire process is essentially automated, requiring minimal human involvement. Intent analysis, automated data optimization, and data derivation rapidly generate candidate expressions, responding more efficiently to business scenario needs and improving timeliness. Historical expression analysis results and relevant experience can be accumulated and reused, avoiding repetitive work and supporting subsequent expression generation and optimization, thus enhancing reusability. Based on precise matching of intent scenarios at perception nodes, expressions tailored to the needs of perception nodes are generated, improving the reach of business support information and achieving personalized expression. Target expressions are evaluated and selected to avoid inefficient expressions consuming resources, improving the overall quality of business support information and optimizing resource utilization.

[0077] In one or more embodiments of this specification, such as Figure 2 As shown, in addition to the large model, the entire system framework also includes infrastructure such as a big data computing platform, neuromorphic computing technology products, a high-performance recall engine, data synchronization / transmission tools, and a cloud-native multi-model database to support the computation of intelligent agents and related content.

[0078] The intelligent agent component system is divided into a scenario configuration layer and a component layer, which are used to configure and use the components, respectively. The scenario configuration layer includes a memory module, tool configuration, and timed triggers. The configuration driver for the memory module and tool configuration enables the unified reuse of intelligent agents in multiple scenarios. The component layer includes a memory component, tools, and triggers. The memory component can record long-term experience, short-term memory, and intelligent agent memory information. The tools include network search, offline analysis, computing tools, report query, etc. The triggers are used to trigger expression analysis, timed interaction, and other content.

[0079] Meanwhile, to ensure the correct citation of historical expression analysis results and related content, a knowledge base can be pre-set to store the relevant content. Typically, at least two databases need to be set up: a business auxiliary information expression knowledge base and a business auxiliary information expression analysis knowledge base.

[0080] The knowledge base for expressing business auxiliary information mainly includes scenario information, online data for expressing business auxiliary information, restrictions on the generation of business auxiliary information, and formats for expressing business auxiliary information. The detailed structure is shown in Table 1 below: Table 1. Structure of the Knowledge Base for Expressing Business Support Information

[0081] Once a batch of business auxiliary information expressions has been produced and deployed in this scenario, and evaluation has been conducted (e.g., A / B testing), business auxiliary information expression analysis can be performed periodically to obtain data for 1 day, 7 days, and 30 days. This analysis will generate an analysis report, which will be distributed to both the business auxiliary information expression analysis report platform and the context engine, resulting in the business auxiliary information expression analysis knowledge base table shown in Table 2. Table 2. Knowledge Base Structure Table for Business Auxiliary Information Expression and Analysis

[0082] To maintain clarity and conciseness, some content in Tables 1 and 2 above has been translated into Chinese (e.g., the descriptions of business auxiliary information identifiers and business auxiliary information names in the case study have been translated into Chinese) and obfuscated (e.g., the network address of product C, xxx: / / xxx.com, etc.) to ensure the overall technical accuracy of the content.

[0083] Meanwhile, when writing context, the input parameter design can be referenced as shown in Table 3 below: Table 3 Input Parameter Design Table

[0084] Meanwhile, for the design of enhanced retrieval and generated RAG, the scene name can be indexed and built according to the dual-index mode, as shown in Table 3 below: Table 4 RAG Design Table

[0085] Based on the same idea, one or more embodiments of this specification also provide apparatus and devices corresponding to the above methods, such as... Figure 5 , Figure 6 As shown.

[0086] Figure 5 This specification provides a schematic diagram of the structure of a business auxiliary information expression device according to one or more embodiments. The device includes: The first intent analysis module 502 performs intent analysis on the node providing business auxiliary information to obtain a first intent analysis result, and analyzes each historical expression corresponding to the business auxiliary information based on the first intent analysis result to obtain a historical expression analysis result. The second intent analysis module 504 performs intent analysis on the perception nodes of business auxiliary information, obtains the second intent analysis result, and determines the corresponding intent scenario based on the second intent analysis result; The data optimization and derivation module 506, for the intent scenario, optimizes and / or derives data from the historical expressions based on the historical expression analysis results to obtain corresponding candidate expressions; The evaluation module 508 evaluates the candidate expressions and selects the corresponding target expression based on the evaluation results, which is used to express the business auxiliary information to the perception node.

[0087] Optionally, the first intent analysis module 502 performs intent analysis on the node providing business auxiliary information through an intent analysis sub-agent to obtain a first intent analysis result; The main agent arranges tasks based on the first intent analysis results to obtain the execution tasks corresponding to each sub-agent. Based on the execution task, the historical expressions corresponding to the business auxiliary information are obtained, and the historical expressions are classified and analyzed according to the corresponding business indicators through the expression analysis sub-agent to obtain the historical expression analysis results.

[0088] Optionally, the first intent analysis module 502, for the service auxiliary information providing node, performs intent analysis through a general intent analysis sub-agent based on the interaction content with the providing node, and obtains a general intent analysis result; Based on the general intent analysis results, intent analysis is performed by the business auxiliary information-specific intent analysis sub-agent to obtain the first intent analysis result.

[0089] Optionally, the first intent analysis module 502 performs a scenario analysis task corresponding to the business auxiliary information expression scenario based on a timed triggering method; Based on the scenario analysis task, sub-expression scenarios to be analyzed are obtained in each business auxiliary information expression scenario; For the aforementioned sub-expression scenario, the corresponding analysis dimension is determined through the scenario analysis sub-agent, and the corresponding scenario data is obtained based on the analysis dimension to perform scenario analysis; Based on the business requirements corresponding to each sub-expression scenario, the corresponding scenario analysis results are filtered and assembled, and interactive content that provides node output is generated.

[0090] Optionally, the second intent analysis module 504 performs intent analysis on the perception node of business auxiliary information through the intent analysis sub-agent to obtain the second intent analysis result; The main agent arranges tasks based on the second intent analysis results to obtain the execution tasks corresponding to each sub-agent. Based on the execution task, the intent scenario in which the perception node is located is obtained according to the second intent analysis result.

[0091] Optionally, the data optimization and derivation module 506, for the intent scenario, selects a first original expression from the historical expressions that does not meet the business requirements based on the historical expression analysis results; The expression optimization sub-agent optimizes the first original expression to obtain the corresponding first candidate expression.

[0092] Optionally, the data optimization and derivation module 506 generates a corresponding second original expression for the intent scenario by generating a sub-agent through expression generation; By using a crowd understanding sub-agent, a crowd profile matching the perception node is obtained based on the second intent analysis result; By expressing the selection sub-agent, a third original expression matching the perception node is obtained from the second original expression based on the crowd profile; By generating a sub-agent through expression, and performing data derivation based on the third original expression, a corresponding second candidate expression is obtained.

[0093] Optionally, the data optimization and derivation module 506 determines, based on the corresponding business indicators, that there is a traffic fluctuation event in the business scenario corresponding to the business auxiliary information. In response to the traffic fluctuation event, the event-targeted audience matching sub-agent determines the profile of a specified audience whose matching degree with the traffic fluctuation event is higher than a first preset degree. For the aforementioned user profile, the third original expression is obtained by filtering through the second original expression, and a specified candidate expression is derived from the data. Based on the changes in the business indicators of the traffic fluctuation event, the time limit corresponding to the specified candidate expression is dynamically set. Within the specified time limit, if it is determined that the sensing node belongs to a designated population based on the population profile of the sensing node, then the designated candidate expression is directly used as the second candidate expression.

[0094] Optionally, the evaluation module 508 performs internal evaluation of the candidate expressions through an evaluation sub-agent; For the candidate expressions that pass the internal evaluation, the import sub-agent is configured to import them into the business scenario corresponding to the business assistance information. Based on the aforementioned business scenario, the imported candidate expressions are subjected to external evaluation.

[0095] Optionally, if the business assistance information is used to jump to its corresponding business scenario, the evaluation module 508 selects several specified expressions with the highest business indicators from each candidate expression according to the evaluation results, and sorts and represents these several specified expressions according to the business indicators. Obtain the perceived expression of the sensing node for the business scenario corresponding to the business assistance information within the most recent preset time period; For the perceived expression, determine its similarity to each specified expression in the ranking representation, and filter out the specified expressions whose similarity is higher than a second preset level; For the remaining specified expressions, the corresponding specified expression is selected as the target expression according to the sorting representation, and is used to express the business auxiliary information to the sensing node.

[0096] Figure 6 This specification provides a schematic diagram of the structure of a business auxiliary information expression device according to one or more embodiments, the device comprising: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, which, when executed by the at least one processor, enable the at least one processor to: Intent analysis is performed on the nodes that provide business auxiliary information to obtain a first intent analysis result. Based on the first intent analysis result, the historical expressions corresponding to the business auxiliary information are analyzed to obtain a historical expression analysis result. Intent analysis is performed on the perception nodes of business auxiliary information to obtain a second intent analysis result, and the corresponding intent scenario is determined based on the second intent analysis result; For the stated intent scenario, based on the historical expression analysis results, the historical expressions are optimized and / or derived to obtain corresponding candidate expressions; The candidate expressions are evaluated, and the corresponding target expression is selected based on the evaluation results to express the business assistance information to the perception node.

[0097] Based on the same idea, one or more embodiments of this specification also provide a non-volatile computer storage medium corresponding to the above method, storing computer-executable instructions, wherein the computer-executable instructions are configured as follows: Intent analysis is performed on the nodes that provide business auxiliary information to obtain a first intent analysis result. Based on the first intent analysis result, the historical expressions corresponding to the business auxiliary information are analyzed to obtain a historical expression analysis result. Intent analysis is performed on the perception nodes of business auxiliary information to obtain a second intent analysis result, and the corresponding intent scenario is determined based on the second intent analysis result; For the stated intent scenario, based on the historical expression analysis results, the historical expressions are optimized and / or derived to obtain corresponding candidate expressions; The candidate expressions are evaluated, and the corresponding target expression is selected based on the evaluation results to express the business assistance information to the perception node.

[0098] In the 1990s, improvements to a technology could be clearly distinguished as either hardware improvements (e.g., improvements to the circuit structure of diodes, transistors, switches, etc.) or software improvements (improvements to methodology). However, with technological advancements, many methodological improvements today can be considered direct improvements to hardware circuit structures. Designers almost always obtain the corresponding hardware circuit structure by programming the improved methodology into the hardware circuit. Therefore, it cannot be said that a methodological improvement cannot be implemented using hardware physical modules. For example, a Programmable Logic Device (PLD) (such as a Field Programmable Gate Array (FPGA)) is such an integrated circuit whose logic function is determined by the user programming the device. Designers can program and "integrate" a digital system onto a PLD themselves, without needing chip manufacturers to design and manufacture dedicated integrated circuit chips. Furthermore, nowadays, instead of manually manufacturing integrated circuit chips, this programming is mostly implemented using "logic compiler" software. Similar to the software compiler used in program development, the original code before compilation must also be written in a specific programming language, called a Hardware Description Language (HDL). There are many HDLs, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, and RHDL (Ruby Hardware Description Language). Currently, the most commonly used are VHDL (Very-High-Speed ​​Integrated Circuit Hardware Description Language) and Verilog. Those skilled in the art should also understand that by simply performing some logic programming on the method flow using one of these hardware description languages ​​and programming it into an integrated circuit, the hardware circuit implementing the logical method flow can be easily obtained.

[0099] The controller can be implemented in any suitable manner. For example, it can take the form of a microprocessor or processor and a computer-readable medium storing computer-readable program code (e.g., software or firmware) executable by the (micro)processor, logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers, and embedded microcontrollers. Examples of controllers include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicon Labs C8051F320. A memory controller can also be implemented as part of the control logic of the memory. Those skilled in the art will also recognize that, in addition to implementing the controller in purely computer-readable program code form, the same functionality can be achieved by logically programming the method steps to make the controller take the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers. Therefore, such a controller can be considered a hardware component, and the means included therein for implementing various functions can also be considered as structures within the hardware component. Alternatively, the means for implementing various functions can be considered as both software modules implementing the method and structures within the hardware component.

[0100] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, a computer can be, for example, a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email device, game console, tablet computer, wearable device, or any combination of these devices.

[0101] For ease of description, the above devices are described in terms of function, divided into various units. Of course, in implementing this specification, the functions of each unit can be implemented in one or more software and / or hardware.

[0102] Those skilled in the art will understand that the embodiments of this specification can be provided as methods, systems, or computer program products. Therefore, the embodiments of this specification can take the form of entirely hardware embodiments, entirely software embodiments, or embodiments combining software and hardware aspects. Furthermore, the embodiments of this specification can take the form of computer program products implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0103] This specification is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this specification. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0104] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0105] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0106] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0107] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0108] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0109] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0110] This specification can be described in the general context of computer-executable instructions that are executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform a specific task or implement a specific abstract data type. This specification can also be practiced in distributed computing environments, where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0111] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the embodiments of apparatus, devices, and non-volatile computer storage media are basically similar to the method embodiments, so the descriptions are relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0112] The foregoing has described specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0113] The above description is merely one or more embodiments of this specification and is not intended to limit this specification. Various modifications and variations can be made to the one or more embodiments of this specification by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of one or more embodiments of this specification should be included within the scope of the claims of this specification.

Claims

1. A method for expressing business auxiliary information, comprising: Intent analysis is performed on the nodes that provide business auxiliary information to obtain a first intent analysis result. Based on the first intent analysis result, the historical expressions corresponding to the business auxiliary information are analyzed to obtain a historical expression analysis result. Intent analysis is performed on the perception nodes of business auxiliary information to obtain a second intent analysis result, and the corresponding intent scenario is determined based on the second intent analysis result; For the stated intent scenario, based on the historical expression analysis results, the historical expressions are optimized and / or derived to obtain corresponding candidate expressions; The candidate expressions are evaluated, and the corresponding target expression is selected based on the evaluation results to express the business assistance information to the perception node.

2. The method as described in claim 1, wherein intent analysis is performed on the node providing business auxiliary information to obtain a first intent analysis result, and the historical expressions corresponding to the business auxiliary information are analyzed based on the first intent analysis result to obtain a historical expression analysis result, specifically including: For the nodes that provide business auxiliary information, intent analysis is performed by the intent analysis sub-agent to obtain the first intent analysis result; The main agent arranges tasks based on the first intent analysis results to obtain the execution tasks corresponding to each sub-agent. Based on the execution task, the historical expressions corresponding to the business auxiliary information are obtained, and the historical expressions are classified and analyzed according to the corresponding business indicators through the expression analysis sub-agent to obtain the historical expression analysis results.

3. The method as described in claim 2, wherein for the node providing business auxiliary information, intent analysis is performed through an intent analysis sub-agent to obtain a first intent analysis result, specifically including: For the nodes that provide business auxiliary information, based on the interaction content with the providing nodes, intent analysis is performed by a general intent analysis sub-agent to obtain general intent analysis results; Based on the general intent analysis results, intent analysis is performed by the business auxiliary information-specific intent analysis sub-agent to obtain the first intent analysis result.

4. The method as described in claim 3, before performing intent analysis on the node providing business auxiliary information to obtain the first intent analysis result, the method further includes: Based on a timed triggering method, execute scenario analysis tasks corresponding to the business auxiliary information expression scenarios; Based on the scenario analysis task, sub-expression scenarios to be analyzed are obtained in each business auxiliary information expression scenario; For the aforementioned sub-expression scenario, the corresponding analysis dimension is determined through the scenario analysis sub-agent, and the corresponding scenario data is obtained based on the analysis dimension to perform scenario analysis; Based on the business requirements corresponding to each sub-expression scenario, the corresponding scenario analysis results are filtered and assembled, and interactive content that provides node output is generated.

5. The method as described in claim 1, wherein intent analysis is performed on the perception node of business auxiliary information to obtain a second intent analysis result, and the corresponding intent scenario is determined based on the second intent analysis result, specifically including: For the perception nodes of business auxiliary information, the intent analysis sub-agent performs intent analysis to obtain the second intent analysis result; The main agent arranges tasks based on the second intent analysis results to obtain the execution tasks corresponding to each sub-agent. Based on the execution task, the intent scenario in which the perception node is located is obtained according to the second intent analysis result.

6. The method as described in claim 1, for the intent scenario, optimizing the historical expressions based on the historical expression analysis results to obtain corresponding candidate expressions, specifically includes: For the stated intent scenario, based on the historical expression analysis results, the first original expression that does not meet the business requirements is selected from the historical expressions; The expression optimization sub-agent optimizes the first original expression to obtain the corresponding first candidate expression.

7. The method as described in claim 1, wherein, for the intent scenario, based on the historical expression analysis results, data derivation is performed on the historical expressions to obtain corresponding candidate expressions, specifically including: For the aforementioned intent scenario, a corresponding second original expression is generated by generating a sub-agent through expression generation; By using a crowd understanding sub-agent, a crowd profile matching the perception node is obtained based on the second intent analysis result; By expressing the selection sub-agent, a third original expression matching the perception node is obtained from the second original expression based on the crowd profile; By generating a sub-agent through expression, and performing data derivation based on the third original expression, a corresponding second candidate expression is obtained.

8. The method of claim 7, further comprising: For the business scenario corresponding to the aforementioned business auxiliary information, it is determined, based on the corresponding business metrics, that the business scenario has a traffic fluctuation event; In response to the traffic fluctuation event, the event-targeted audience matching sub-agent determines the profile of a specified audience whose matching degree with the traffic fluctuation event is higher than a first preset degree. For the aforementioned user profile, the third original expression is obtained by filtering through the second original expression, and a specified candidate expression is derived from the data. Based on the changes in the business indicators of the traffic fluctuation event, the time limit corresponding to the specified candidate expression is dynamically set. Within the specified time limit, if it is determined that the sensing node belongs to a designated population based on the population profile of the sensing node, then the designated candidate expression is directly used as the second candidate expression.

9. The method of claim 1, wherein evaluating the candidate expression specifically includes: The candidate expressions are internally evaluated by evaluating sub-agents; For the candidate expressions that pass the internal evaluation, the import sub-agent is configured to import them into the business scenario corresponding to the business assistance information. Based on the aforementioned business scenario, the imported candidate expressions are subjected to external evaluation.

10. The method as described in claim 1, wherein a corresponding target representation is selected based on the evaluation result for expressing the business assistance information to the sensing node, specifically including: If the business assistance information is used to jump to its corresponding business scenario, then according to the evaluation results, select a number of designated expressions with the highest business indicators from each candidate expression, and sort and represent these designated expressions according to the business indicators; Obtain the perceived expression of the sensing node for the business scenario corresponding to the business assistance information within the most recent preset time period; For the perceived expression, determine its similarity to each specified expression in the ranking representation, and filter out the specified expressions whose similarity is higher than a second preset level; For the remaining specified expressions, the corresponding specified expression is selected as the target expression according to the sorting representation, and is used to express the business auxiliary information to the sensing node.

11. A business auxiliary information expression device, comprising: The first intent analysis module performs intent analysis on the node providing business auxiliary information to obtain the first intent analysis result, and analyzes each historical expression corresponding to the business auxiliary information based on the first intent analysis result to obtain the historical expression analysis result. The second intent analysis module performs intent analysis on the perception nodes of business auxiliary information, obtains the second intent analysis result, and determines the corresponding intent scenario based on the second intent analysis result; The data optimization and derivation module, for the intent scenario, optimizes and / or derives data from the historical expressions based on the historical expression analysis results to obtain corresponding candidate expressions; The evaluation module evaluates the candidate expressions and selects the corresponding target expression based on the evaluation results, which is used to express the business auxiliary information to the perception node.

12. The apparatus of claim 11, wherein the first intent analysis module performs intent analysis on the node providing business auxiliary information through an intent analysis sub-agent to obtain a first intent analysis result; The main agent arranges tasks based on the first intent analysis results to obtain the execution tasks corresponding to each sub-agent. Based on the execution task, the historical expressions corresponding to the business auxiliary information are obtained, and the historical expressions are classified and analyzed according to the corresponding business indicators through the expression analysis sub-agent to obtain the historical expression analysis results.

13. The apparatus of claim 12, wherein the first intent analysis module, for the node providing business auxiliary information, performs intent analysis through a general intent analysis sub-agent based on the interaction content with the node providing the business auxiliary information, and obtains a general intent analysis result; Based on the general intent analysis results, intent analysis is performed by the business auxiliary information-specific intent analysis sub-agent to obtain the first intent analysis result.

14. The apparatus of claim 13, wherein the first intent analysis module performs a scenario analysis task corresponding to the business auxiliary information expression scenario based on a timed triggering method; Based on the scenario analysis task, sub-expression scenarios to be analyzed are obtained in each business auxiliary information expression scenario; For the aforementioned sub-expression scenario, the corresponding analysis dimension is determined through the scenario analysis sub-agent, and the corresponding scenario data is obtained based on the analysis dimension to perform scenario analysis; Based on the business requirements corresponding to each sub-expression scenario, the corresponding scenario analysis results are filtered and assembled, and interactive content that provides node output is generated.

15. The apparatus of claim 11, wherein the second intent analysis module performs intent analysis on the sensing node of business auxiliary information through an intent analysis sub-agent to obtain a second intent analysis result; The main agent arranges tasks based on the second intent analysis results to obtain the execution tasks corresponding to each sub-agent. Based on the execution task, the intent scenario in which the perception node is located is obtained according to the second intent analysis result.

16. The apparatus of claim 11, wherein the data optimization and derivation module, for the intent scenario, selects a first original expression from the historical expressions that does not meet the business requirements based on the historical expression analysis results; The expression optimization sub-agent optimizes the first original expression to obtain the corresponding first candidate expression.

17. The apparatus of claim 11, wherein the data optimization and derivation module generates a corresponding second original expression for the intent scenario by generating a sub-agent through expression generation; By using a crowd understanding sub-agent, a crowd profile matching the perception node is obtained based on the second intent analysis result; By expressing the selection sub-agent, a third original expression matching the perception node is obtained from the second original expression based on the crowd profile; By generating a sub-agent through expression, and performing data derivation based on the third original expression, a corresponding second candidate expression is obtained.

18. The apparatus of claim 17, wherein the data optimization and derivation module determines, based on corresponding business indicators, that a traffic fluctuation event exists in the business scenario corresponding to the business auxiliary information; In response to the traffic fluctuation event, the event-targeted audience matching sub-agent determines the profile of a specified audience whose matching degree with the traffic fluctuation event is higher than a first preset degree. For the aforementioned user profile, the third original expression is obtained by filtering through the second original expression, and a specified candidate expression is derived from the data. Based on the changes in the business indicators of the traffic fluctuation event, the time limit corresponding to the specified candidate expression is dynamically set. Within the specified time limit, if it is determined that the sensing node belongs to a designated population based on the population profile of the sensing node, then the designated candidate expression is directly used as the second candidate expression.

19. The apparatus of claim 11, wherein the evaluation module performs internal evaluation of the candidate expression through an evaluation sub-agent; For the candidate expressions that pass the internal evaluation, the import sub-agent is configured to import them into the business scenario corresponding to the business assistance information. Based on the aforementioned business scenario, the imported candidate expressions are subjected to external evaluation.

20. The apparatus of claim 11, wherein if the business assistance information is used to jump to its corresponding business scenario, the evaluation module selects several specified expressions with the highest business indicators from each candidate expression according to the evaluation results, and sorts and represents the several specified expressions according to the business indicators; Obtain the perceived expression of the sensing node for the business scenario corresponding to the business assistance information within the most recent preset time period; For the perceived expression, determine its similarity to each specified expression in the ranking representation, and filter out the specified expressions whose similarity is higher than a second preset level; For the remaining specified expressions, the corresponding specified expression is selected as the target expression according to the sorting representation, and is used to express the business auxiliary information to the sensing node.

21. A business-aided information expression device, comprising: At least one processor; as well as, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, which, when executed by the at least one processor, enable the at least one processor to: Intent analysis is performed on the nodes that provide business auxiliary information to obtain a first intent analysis result. Based on the first intent analysis result, the historical expressions corresponding to the business auxiliary information are analyzed to obtain a historical expression analysis result. Intent analysis is performed on the perception nodes of business auxiliary information to obtain a second intent analysis result, and the corresponding intent scenario is determined based on the second intent analysis result; For the stated intent scenario, based on the historical expression analysis results, the historical expressions are optimized and / or derived to obtain corresponding candidate expressions; The candidate expressions are evaluated, and the corresponding target expression is selected based on the evaluation results to express the business assistance information to the perception node.

Images