Digital service session interaction processing method and system

Abstract

The invention provides a digital service session interaction processing method and system, and relates to the technical field of digital service session interaction process.The digital service session interaction processing method comprises the steps that firstly, digital service session source data are obtained, and session interaction gene segments bearing core semantics and function association are extracted from the digital service session source data; constructing a digital service session interaction gene chain which comprises a plurality of gene nodes and associated links based on the session interaction gene segment, extracting a user instant session intention according to the gene chain and the current input content, generating an intention symbiotic evolution unit, receiving a user feedback text, analyzing semantic deviation and correcting a gene chain evolution path; according to the method, a gene node association relationship and intent symbiotic evolution unit content are updated according to a correction path, an iterated gene chain is obtained, an evolution rule and an adaptation mode are summarized finally, a digital service session interaction standard evolution mechanism for guiding subsequent interaction is output, the accuracy of digital service session interaction is greatly improved, and the interaction efficiency is improved. And the user experience and the digital service quality are obviously improved.

Description

Technical Field

[0001] This invention relates to the field of digital service session interaction processing technology, and more specifically, to a digital service session interaction processing method and system. Background Technology

[0002] In today's era of booming digital services, digital service conversational interaction has become the primary way for users to communicate with digital service systems. Existing digital service conversational interaction processing technologies typically focus on performing simple semantic analysis on the text input by users to identify their basic needs and match corresponding digital service functions.

[0003] However, these technologies have significant limitations. On the one hand, they often treat the user's historical conversation text, current conversation input, and digital service function description text in isolation, lacking an exploration of the deep semantic and functional relationships between the three. For example, when processing multi-turn conversations, it is difficult to effectively capture the inheritance and changes of core semantics across different turns, resulting in an inaccurate understanding of user intent.

[0004] On the other hand, existing technologies lack the ability to dynamically adjust session interaction processing strategies when responding to user feedback. When user feedback does not match system expectations, the evolution path of session interaction cannot be corrected in a timely manner, making it difficult for subsequent interactions to effectively meet user needs and affecting the quality of digital services and user experience. Summary of the Invention

[0005] In view of the aforementioned problems, and in conjunction with the first aspect of the present invention, embodiments of the present invention provide a digital service session interaction processing method, the method comprising: Obtain digital service session source data, and extract session interaction gene fragments from the digital service session source data. The digital service session source data includes user historical session text, current session input content, and digital service function description text. The session interaction gene fragments refer to text units that carry the core semantics and functional associations of the session. Based on the aforementioned conversational interaction gene fragments, the semantic inheritance relationship and functional adaptation relationship between each conversational interaction gene fragment are associated to construct a digital service conversational interaction gene chain, which includes multiple gene nodes and the associated links between gene nodes. Based on the digital service session interaction gene chain and the user's current session input, the user's real-time session intent is extracted, and the user's real-time session intent is associated with the functional association information in the digital service session interaction gene chain to generate an intent co-evolution unit. Receive user feedback text for the intent co-evolution unit, analyze the semantic deviation between the feedback text and the intent co-evolution unit, and correct the evolution path of the digital service session interaction gene chain based on the semantic deviation to obtain the corrected evolution path. According to the modified evolution path, update the gene node association relationship and the content of the intention co-evolution unit in the digital service session interaction gene chain to obtain the iterative digital service session interaction gene chain. Based on the iterative digital service session interaction gene chain, the evolution law and intent symbiosis adaptation mode of the digital service session interaction gene chain are summarized, and a standard evolution mechanism for digital service session interaction is output. The standard evolution mechanism for digital service session interaction is used to guide the construction of the digital service session interaction gene chain and the intent symbiosis process of subsequent session interactions.

[0006] Furthermore, embodiments of the present invention also provide a digital service session interaction processing system, characterized in that it includes: A processor; a machine-readable storage medium for storing machine-executable instructions of the processor; wherein the processor is configured to perform the above-described digital service session interaction processing method by executing the machine-executable instructions.

[0007] In another aspect, embodiments of the present invention also provide a computer program product, the computer program product including machine-executable instructions stored in a computer-readable storage medium, a processor of a computer device reading the machine-executable instructions from the computer-readable storage medium, and the processor executing the machine-executable instructions to cause the computer device to perform the above-described digital service session interaction processing method.

[0008] Based on the above, by accurately extracting conversation interaction gene fragments from digital service conversation source data, we can deeply mine text units that carry the core semantics and functional associations of the conversation. Based on these gene fragments, we construct a digital service conversation interaction gene chain, linking the semantic inheritance and functional adaptation relationships between each gene fragment. This achieves structured modeling of the conversation interaction process. After extracting the user's immediate conversation intent and generating intent symbiotic evolution units, we receive user feedback and analyze semantic deviations, thereby correcting the evolution path of the digital service conversation interaction gene chain. This demonstrates the system's dynamic adaptability and self-optimization capabilities, enabling timely adjustments to interaction strategies based on actual user feedback to ensure subsequent interactions better meet user needs. Finally, by iteratively updating the digital service conversation interaction gene chain, we summarize evolutionary patterns and intent symbiotic adaptation models, outputting a standard evolution mechanism for digital service conversation interaction. This provides norms and guidance for subsequent conversation interactions, significantly improving the accuracy, efficiency, and intelligence of digital service conversation interactions, and substantially enhancing user experience and the quality of digital services. Attached Figure Description

[0009] Figure 1 This is a schematic diagram of the execution flow of the digital service session interaction processing method provided in the embodiments of the present invention.

[0010] Figure 2 This is a schematic diagram of exemplary hardware and software components of the digital service session interaction processing system provided in an embodiment of the present invention. Detailed Implementation

[0011] The present invention will now be described in detail with reference to the accompanying drawings. Figure 1 This is a flowchart illustrating a digital service session interaction processing method according to an embodiment of the present invention. The following is a detailed description of the digital service session interaction processing method.

[0012] Step S110: Obtain digital service session source data, and extract session interaction gene fragments from the digital service session source data. The digital service session source data includes user historical session text, current session input content, and digital service function description text. The session interaction gene fragments refer to text units that carry the core semantics and functional associations of the session.

[0013] In this embodiment, the acquisition of digital service session source data requires a comprehensive approach across multiple channels. Historical user session text can be retrieved from a long-running session storage system. This system stores past conversations between the user and the digital service in a structured format, covering the time sequence of the conversations, the business topics involved, and the roles of the interacting parties. Current session input content is obtained by real-time monitoring of the user's interaction interface with the digital service. This interface supports multiple input formats, including text and voice, and the input content is instantly converted into text format for subsequent processing. Digital service function description text is extracted from the server's function document library, which contains detailed descriptions of each digital service function, including the function's business logic, triggering conditions, interaction flow, and data constraints. After acquisition, the above data is processed to extract session interaction gene fragments.

[0014] Step S111: Separate the user's historical session text, current session input content, and digital service function description text from the digital service session source data, and divide the data source partitions corresponding to the user's historical session text, current session input content, and digital service function description text respectively.

[0015] In this embodiment, the acquired digital service session source data is separated based on its source and attributes. User historical session text is extracted from the session storage system via database queries. During extraction, it is initially categorized according to the business type of the dialogue (e.g., consultation, transaction, complaint, etc.) and then marked as a user historical session text partition. Current session input content is captured by the interactive interface's listening module. After capture, it is marked according to the input's business scenario (e.g., service consultation scenario, business processing scenario, etc.) and then marked as a current session input content partition. Digital service function description text is obtained by accessing the function document library's storage path. After acquisition, it is categorized according to the function's module (e.g., user management module, business processing module, data statistics module, etc.) and marked as a digital service function description text partition. Through these operations, different types of data are clearly distinguished.

[0016] Step S112: Cross-compare the text content of the digital service session source data in each partition and remove duplicate text content.

[0017] In this embodiment, the text content of the three partitions is cross-referenced pairwise. Specifically, the text content of each partition is converted into a string sequence, and then a deep learning-based text similarity matching model (such as a BERT-based text similarity model) is used to compare the string sequences of different partitions. For example, in the user's historical conversation text, there may be statements such as "What is the business process for digital services? What conditions need to be met? What materials need to be submitted?", and the digital service function description text also contains the same description of "business process, processing conditions, and required materials". When the semantic similarity of these text contents reaches a preset threshold (such as 0.9 or higher), only one of them is retained. During the comparison process, the semantic similarity of the text and the consistency of business logic need to be considered. Text content with the same semantics but slightly different expression and format is also judged as duplicate content and removed to ensure data simplicity and avoid repeated analysis of the same content in subsequent processing. At the same time, for some text containing dynamic information (such as time, number, etc.), the dynamic information needs to be standardized before comparison.

[0018] Step S113: Scan the user's historical conversation text segment by segment to identify the paragraphs in the user's historical conversation text that contain the user's core needs statement and digital service response semantics. The user's core needs statement refers to the core demands of the user for digital services expressed in the conversation, and the digital service response semantics refers to the functional related responses given by the digital service to the user's needs.

[0019] Step S1131: Divide the user's historical conversation text into multiple consecutive conversation segments according to the time sequence of the conversation. Each conversation segment contains a complete set of user input and digital service response interactions.

[0020] In this embodiment, the text is segmented according to the time sequence and business process nodes based on the timestamp information of the dialogue in the user's historical conversation text and the business interaction logic. For example, at time T1, a user enters, "I need to upgrade my digital services. What are the specific steps? What benefits will change after the upgrade? Will it affect my existing services?" Then, at time T2, the digital service responds, "Hello, the steps for upgrading your services are as follows: First, log in to the digital service platform, go to the 'Service Management' module, and find the 'Service Upgrade' entry. Then, select the service type and upgrade package you want to upgrade according to the page prompts. Next, confirm the upgrade information and submit it. The system will automatically verify the service. After verification, you need to complete the corresponding payment (if any). Finally, after payment, the system will complete the service upgrade configuration within 1-3 business days. You can view the upgraded service status in 'My Services'. Benefits after the upgrade include: service bandwidth increased to twice the original, added data backup service, and priority technical support. The upgrade process will not affect your existing services, but some new features may not be available immediately during the configuration period. Once the configuration is complete, all functions will be available normally." This user input and the digital service response constitute a conversation segment. If a user enters "How much does it cost to upgrade the service? Can I use coupons? Can I downgrade if I'm not satisfied with the upgrade?" at time T3, the digital server responds at time T4 with "Hello, the cost of upgrading the service depends on the upgrade package you choose. The specific cost will be displayed on the page when you select the package. You can use the general coupons issued by the platform, but some exclusive packages may not support coupon usage. Please refer to the page prompts for details. If you are not satisfied with the upgrade, you can apply for a downgrade within 7 days of the upgrade. When downgrading, you will need to pay or receive a refund according to the downgraded service package (depending on the package price difference). The service benefits after downgrading will be implemented according to the new package standard, and the downgrade operation will be configured within 1-2 business days." This constitutes a new conversation segment. In this way, the user's historical conversation text is divided into multiple conversation segments containing complete interactive content, with each segment corresponding to the process of raising and resolving a business issue.

[0021] Step S1132: Perform semantic parsing on each conversation segment and extract the core vocabulary of the user input statement in the conversation segment. The core vocabulary refers to nouns, verbs or adjectives that can represent the core demands of the user input statement.

[0022] In this embodiment, a semantic parsing model based on the Transformer architecture (such as the BERT model) is used to parse the user input statements in each conversation segment. For example, for the user input "I need to upgrade my digital services. What are the specific steps? What benefits will change after the upgrade? Will it affect my existing service usage?", semantic parsing extracts the core words "service upgrade", "steps", "benefits changes", and "impact on service usage". For the user input "How much does the service upgrade cost? Can I use coupons? If I'm not satisfied with the upgrade, can I downgrade?", the core words extracted are "service upgrade cost", "coupon usage", and "downgrade". When extracting core words, it is necessary to filter out some meaningless auxiliary words and modifiers, and only retain nouns, verbs, or adjectives that can represent the user's core needs. At the same time, it is necessary to ensure that the extracted core words can accurately reflect the user's needs.

[0023] Step S1133: Based on the extracted core words, determine whether the user input statement contains a definite demand, and mark the conversation paragraph containing the user input statement containing the definite demand as the demand-related paragraph.

[0024] In this embodiment, the semantic meaning of the extracted core words is used to determine whether the user's input statement has a clear need. For example, if the core words are "business upgrade," "operation steps," "changes in rights and interests," or "impact of business usage," it indicates that the user has a clear need related to business upgrades, and the conversation paragraph containing this user's input statement is marked as a need-related paragraph. If the user inputs "just asking about digital services," the extracted core words do not reflect a clear need, and the conversation paragraph is not marked as a need-related paragraph. During the judgment process, a preset need-related vocabulary library can be referenced. When the extracted core words match the words in the vocabulary library, it is determined that it contains a definite need. At the same time, for some user input statements with implicit needs, it is necessary to analyze them in conjunction with the business scenario and contextual information to accurately determine whether they contain a definite need.

[0025] Step S1134: Parse the digital service response statements in the requirement-related paragraphs, and extract the functional references in the digital service response statements that are related to the user's requirements. The functional references refer to the descriptions of the functional modules involved in the digital service response that can meet the user's requirements.

[0026] In this embodiment, a semantic parsing model based on the Transformer architecture is also used to parse the digital service response statements in the request-related paragraph. For example, in the request-related paragraph, the user enters, "I need to upgrade my digital service. What are the specific steps? What changes will occur to my benefits after the upgrade? Will it affect my existing service usage?" The digital service responds, "Hello, the steps for upgrading your service are as follows: First, you need to log in to the digital service platform, go to the 'Service Management' module, and find the 'Service Upgrade' entry; then, select the service type and upgrade package you want to upgrade according to the page prompts; next, confirm the upgrade information and submit it. The system will automatically verify the service. After the verification is successful, you need to complete the corresponding payment (if there is a fee); finally, after payment, the system will complete the service upgrade configuration within 1-3 business days." You can view the upgraded service status in 'My Services'. Upgraded benefits include: doubled service bandwidth, added data backup service, and priority technical support. The upgrade will not affect your existing services; however, some new features may not be immediately available during configuration. All features will be available after configuration is complete. Semantic parsing extracts relevant user-required content such as "Service Management Module," "Service Upgrade Entry," "Service Type Selection," "Package Selection," "Service Verification," "Payment," "Service Upgrade Configuration," "My Services Module," "Service Bandwidth Increase," "Data Backup Service," "Priority Technical Support," and "Impact on Service Usage." During extraction, it's crucial to ensure that the extracted content directly relates to user needs, excluding irrelevant information such as greetings and interjections, and accurately identifying the module names, operation steps, and benefits.

[0027] Step S1135: Determine whether the function references in the digital service response statement semantically correspond to the needs indicated by the user input statement, and confirm that the semantically corresponding conversation paragraph contains digital service response semantics.

[0028] In this embodiment, the extracted user input statement's core keywords related to the user's needs are semantically matched with the functional references in the digital service response statement. For example, if the user's needs are "business upgrade," "operation steps," "changes in rights," or "impact of business usage," and the functional references are "business management module," "business upgrade entry," "business type selection," "package selection," "business verification," "payment," "business upgrade configuration," "my business module," "service bandwidth improvement," "data backup service," "priority technical support," or "explanation of the impact of business usage," and the two semantics correspond, then the conversation segment contains digital service response semantics. If the functional references are unrelated to the user's needs, such as the user needing a business upgrade but the functional references being user registration, then the conversation segment does not contain digital service response semantics. During the judgment process, semantic similarity calculation methods can be used, such as cosine similarity calculation based on word vectors. When the similarity reaches a preset threshold (e.g., above 0.8), it is determined that a semantic correspondence has been formed. Furthermore, for some complex business scenarios, a comprehensive judgment combining business logic and contextual information is necessary to ensure the accuracy of the semantic correspondence.

[0029] Step S1136: Summarize all conversation paragraphs containing user core requirement statements and digital service response semantics to obtain paragraphs containing user requirement statements and corresponding digital service response semantics.

[0030] In this embodiment, the conversation segments containing the core user requirements and digital service response semantics confirmed in step S1135 are summarized into a set. For example, after the preceding steps, multiple such conversation segments are obtained, and they are organized together to facilitate the extraction of text fragments from these segments later. During the summarization process, the conversation segments need to be numbered and indexed for subsequent management and querying.

[0031] Step S114: Extract text segments from the identified paragraphs that can independently carry a single semantic and functional association. The single semantic and functional association refers to the text segment containing only one user demand and the corresponding digital service function association information.

[0032] In this embodiment, text fragments are extracted from the paragraphs summarizing the core user needs and digital service response semantics obtained in step S1136. For example, in a conversation paragraph, a user says, "I need to upgrade my digital service. What are the specific steps?", and the digital service responds, "Hello, the steps for upgrading your service are as follows: First, you need to log in to the digital service platform, enter the 'Business Management' module, and find the 'Business Upgrade' entry; then, select the service type and upgrade package you want to upgrade according to the page prompts; next, confirm the upgrade information and submit it. The system will automatically verify the service. After the verification is successful, you need to complete the corresponding payment (if there is a fee); finally, after payment, the system will complete the service upgrade configuration within 1-3 business days. You can view the upgraded service status in 'My Services'." Therefore, the text fragment "I need to upgrade my digital service. What are the specific steps?" can be extracted as a text fragment. This text fragment contains a user need for information on the steps for upgrading the service. The corresponding digital service function association information includes the Business Management module, Business Upgrade Entry, Business Type Selection, Package Selection, Business Verification, Payment, Business Upgrade Configuration, and My Business module. Alternatively, you can extract the following text: "Hello, the steps for business upgrade are as follows: First, you need to log in to the digital service platform, enter the 'Business Management' module, and find the 'Business Upgrade' entry; then, select the business type and upgrade package you want to upgrade according to the page prompts; next, confirm the upgrade information and submit it. The system will automatically perform business verification. After verification, you need to complete the corresponding payment (if any); finally, after payment, the system will complete the business upgrade configuration within 1-3 business days. You can view the upgraded business status in 'My Business.'" This text fragment indicates that the user wants to upgrade their business, and the corresponding digital service function association information is the above-mentioned operation steps. During the extraction process, it is necessary to ensure that each text fragment contains only one user need and its corresponding digital service function association information to avoid multiple needs or mixed function association information. At the same time, for some longer text content, it is necessary to reasonably truncate it according to the semantic integrity and the independence of business logic, so as to ensure that the truncated text fragments can accurately carry a single semantic and functional association.

[0033] Step S115: Scan the digital service function description text and extract the text content describing the core function, applicable scenarios, and interaction methods of the function.

[0034] Step S1151: Analyze the digital service function description text sentence by sentence and identify the sentences containing function names in the digital service function description text.

[0035] In this embodiment, a Transformer-based syntactic analysis model (such as the spaCy syntactic analysis model) is used to parse the digital service function description text sentence by sentence. For example, in the digital service function description text, there is a sentence like "Business upgrade function: Users can upgrade their existing digital service services to improve service benefits and user experience." Syntactic analysis identifies the statement containing the function name "Business upgrade function." Similarly, in the sentence "Data backup function: Users can regularly back up their business data to prevent data loss," the statement containing the function name "Data backup function" is identified. During the identification process, it is necessary to accurately identify the function name, including the official name and common aliases, to ensure accurate subsequent function partitioning. Furthermore, for some compound function names, it is necessary to analyze them in conjunction with contextual information to ensure the accuracy of function name identification.

[0036] Step S1152: Based on the identified function names, divide the digital service function description text into function partitions, with each function partition corresponding to a digital service function module.

[0037] In this embodiment, the text describing digital service functions is divided into different functional partitions based on the identified function names. For example, using "business upgrade function" as a benchmark, text containing descriptions related to this function is divided into a business upgrade function partition; using "data backup function" as a benchmark, text containing descriptions related to this function is divided into a data backup function partition. During the partitioning process, it is necessary to ensure that the text content of each functional partition is related to the corresponding function name to avoid content mixing across functional partitions. Furthermore, for text content involving multiple functions, it is necessary to reasonably partition it according to the business logic relationship between the functions to ensure the accuracy and rationality of the functional partitioning.

[0038] Step S1153: Within each functional partition, extract the text content describing the core function of the functional module. The core function refers to the core problem that the functional module can solve for the user or the core value it can provide.

[0039] In this embodiment, the core function text is extracted from the text content of each functional partition. For example, in the business upgrade function partition, the text content is "Business upgrade function: Users can upgrade their existing digital service services to improve service benefits and user experience, solve the problem of insufficient business functions failing to meet user needs, and provide users with higher quality and more efficient digital services." Through semantic analysis, the text content describing the core function is extracted as "Users can upgrade their existing digital service services to improve service benefits and user experience, solve the problem of insufficient business functions failing to meet user needs, and provide users with higher quality and more efficient digital services." In the data backup function partition, the text content is "Data backup function: Users can regularly back up their business data to prevent data loss, ensure the security and integrity of user data, and solve the problem of business interruption and loss caused by data loss." The text content describing the core function is extracted as "Users can regularly back up their business data to prevent data loss, ensure the security and integrity of user data, and solve the problem of business interruption and loss caused by data loss." During the extraction process, it is necessary to accurately identify the core problems that functional modules can solve for users or the core value they can provide, and exclude some irrelevant descriptions, such as the technical implementation details of the functions.

[0040] Step S1154: Within each functional partition, extract the text content describing the applicable scenarios of the functional module. The applicable scenarios refer to the condition descriptions of the scenarios in which the functional module can function normally and meet the user's needs.

[0041] In this embodiment, the applicable scenario text is extracted for the text content of each functional partition. For example, in the business upgrade function partition, the text content is "Business upgrade function: Users can upgrade their existing digital service services to improve service benefits and user experience. This is suitable for users who have used the service for a long time, have higher requirements for service quality, or users whose current needs cannot be met by the service function." The extracted text content describing the applicable scenario is "Suitable for users who have used the service for a long time, have higher requirements for service quality, or users whose current needs cannot be met by the service function." In the data backup function partition, the text content is "Data backup function: Users can regularly back up their business data to prevent data loss. This is suitable for users whose business data is highly important and whose data is updated frequently, or users who have strict requirements for data security." The extracted text content describing the applicable scenario is "Suitable for users whose business data is highly important and whose data is updated frequently, or users who have strict requirements for data security." During the extraction process, it is necessary to accurately identify the conditions that allow the functional module to function normally and meet the user's needs, including conditions such as user type, business scenario, and data characteristics, to ensure that the extracted applicable scenario text accurately reflects the usage conditions of the function.

[0042] Step S1155: Within each functional area, extract the text content describing the interaction method of the functional module. The functional interaction method refers to the description of the input form, operation steps, and response form required when the user operates the functional module.

[0043] In this embodiment, the text describing the interaction method is extracted for the text content of each functional area. For example, in the business upgrade functional area, the text content is: "Business upgrade function: Users can log in to the digital service platform, enter the 'Business Management' module, find the 'Business Upgrade' entry, then select the business type and upgrade package according to the page prompts, confirm the information and submit, and complete the payment (if any). After that, the system will complete the business upgrade configuration within 1-3 working days, and users can view the upgraded business status in 'My Business'." The extracted text describes the interaction method: "Users can log in to the digital service platform, enter the 'Business Management' module, find the 'Business Upgrade' entry, then select the business type and upgrade package according to the page prompts, confirm the information and submit, and complete the payment (if any). After that, the system will complete the business upgrade configuration within 1-3 working days, and users can view the upgraded business status in 'My Business'." Users can view the upgraded business status in 'My Business'; in the data backup function section, the text content is 'Data Backup Function: Users can log in to the digital service platform, enter the 'Data Management' module, find the 'Data Backup' entry, then select the backup cycle and backup method (full backup or incremental backup), confirm the backup information and submit. The system will automatically perform data backup at the specified time and send a notification to the user after the backup is completed.' The text describing the interaction method is extracted as follows: 'Users can log in to the digital service platform, enter the 'Data Management' module, find the 'Data Backup' entry, then select the backup cycle and backup method (full backup or incremental backup), confirm the backup information and submit. The system will automatically perform data backup at the specified time and send a notification to the user after the backup is completed.' During the extraction process, it is necessary to accurately identify the input forms (such as login, selection, submission, etc.), operation steps (such as entering the module, finding the entry, selecting parameters, etc.), and response forms (such as system configuration, sending notifications, etc.) required for users to operate this function module, to ensure that the extracted interaction method text accurately reflects the function's operation flow.

[0044] Step S1156: Integrate the text content extracted from each functional area, which describes the core function, applicable scenarios, and interaction methods of the function, into a functional description unit. Each functional description unit corresponds to the complete association information of a functional module.

[0045] In this embodiment, for each functional area, the extracted text content describing the core function, applicable scenarios, and interaction methods is integrated. For example, in the business upgrade function area, the integrated function description unit is: "Business Upgrade Function: The core function is that users can upgrade their existing digital service services to improve service benefits and user experience, solve the problem of insufficient business functions failing to meet user needs, and provide users with higher quality and more efficient digital services; applicable scenarios are suitable for users who have used the service for a long time, have higher requirements for service quality, or whose current needs cannot be met by the business functions; the interaction method is that users can log in to the digital service platform, enter the 'Business Management' module, find the 'Business Upgrade' entry, then select the business type and upgrade package according to the page prompts, confirm the information and submit, and complete the payment (if any). The system will complete the business upgrade configuration within 1-3 working days, and users can check their 'My' account." View the upgraded business status in the "Business" section. In the data backup function section, the integrated function description unit is: "Data Backup Function: Its core function is to allow users to regularly back up their business data, preventing data loss, ensuring the security and integrity of user data, and resolving business interruptions and losses caused by data loss. It is suitable for users with highly important business data, frequent data updates, or those with strict data security requirements. The interaction method is for users to log in to the digital service platform, enter the 'Data Management' module, find the 'Data Backup' entry, select the backup cycle and backup method (full backup or incremental backup), confirm the backup information and submit. The system will automatically perform data backup at the specified time and send a notification to the user upon completion." During the integration process, it is necessary to ensure that the content of the function description unit is complete, logically clear, and accurately reflects the complete association information of a functional module.

[0046] Step S1157: Analyze the semantic length and expression style of user needs in the user's historical conversation text. Based on the semantic length and expression style of user needs in the user's historical conversation text, the functional description unit is divided into multiple short text fragments. The semantic length of each short text fragment is similar to the semantic length of user needs in the user's historical conversation text.

[0047] In this embodiment, the semantic length of user requests expressed in historical user conversations is first statistically analyzed to calculate the average semantic length (in characters or words). Simultaneously, the expression style of user requests is analyzed (e.g., concise and clear, detailed description, interrogative tone, etc.). For example, statistics show that the average semantic length of user requests expressed in historical user conversations is 50-80 characters, and the expression style is mostly concise and clear interrogative or declarative sentences. Then, based on the analyzed semantic length and expression style, the functional description unit is divided into multiple short text fragments. For example, the functional description unit of the business upgrade function is broken down into short text fragments such as "Business upgrade function: Users can upgrade digital service services," "Business upgrade function: Improve service benefits and user experience," "Business upgrade function: Solve the problem of insufficient business functions," "Business upgrade function: Suitable for users who have been using the service for a long time," "Business upgrade function: Suitable for users whose needs are not met," "Business upgrade function: Log in to the platform to enter the business management module," "Business upgrade function: Select business type and package," "Business upgrade function: Upgrade configuration after completing payment," and "Business upgrade function: View status in My Services." Each short text fragment has a semantic length of about 50-80 characters, and the expression style is similar to the user's requirement expression. In the process of decomposition, it is necessary to ensure that the semantics of each short text fragment are complete and can independently carry the information related to a function, while the semantic length is similar to the semantic length of the user's requirement expression, so as to ensure that the subsequent semantic comparison can be carried out smoothly.

[0048] Step S1158: Semantically annotate the split short text fragments, annotating the functional attributes corresponding to each short text fragment. The functional attributes include core function attributes, applicable scenario attributes, or interaction method attributes, to obtain functionally associated fragments that can be associated with the user's needs.

[0049] In this embodiment, a rule-based semantic annotation method or a deep learning-based semantic annotation model is used to semantically annotate the segmented short text fragments. For example, for the short text fragment "Business upgrade function: Users can upgrade digital service business", its corresponding functional attribute is determined to be a core function attribute based on its content, and it is annotated as "Core function attribute - Business upgrade function - Users can upgrade digital service business"; for the short text fragment "Business upgrade function: Applicable to users who have used the business for a long time", its corresponding functional attribute is determined to be an applicable scenario attribute, and it is annotated as "Applicable scenario attribute - Business upgrade function - Applicable to users who have used the business for a long time"; for the short text fragment "Business upgrade function: Log in to the platform to enter the business management module", its corresponding functional attribute is determined to be an interaction method attribute, and it is annotated as "Interaction method attribute - Business upgrade function - Log in to the platform to enter the business management module". During the annotation process, it is necessary to ensure that the annotation of functional attributes is accurate and can reflect the semantic content of the short text fragments so as to semantically associate it with the user's needs in the future.

[0050] Step S116: Perform semantic comparison between the text fragments extracted from the user's historical conversation text and the function-related fragments extracted from the digital service function description text, retain the fragments with semantic matching degree that meet the preset requirements, and exclude the fragments with semantic irrelevance.

[0051] In this embodiment, a deep learning-based text semantic matching model (such as the BERT-based text semantic matching model) is used to semantically compare text fragments extracted from the user's historical conversation text with functionally related fragments extracted from the digital service function description text. For example, the text fragment extracted from the user's historical conversation text might be "I need to upgrade my digital service business. What are the specific steps?", while the functionally related fragments extracted from the digital service function description text might be "Business upgrade function: Log in to the platform to enter the business management module," "Business upgrade function: Select business type and package," "Business upgrade function: Upgrade configuration after payment," "Business upgrade function: View status in my business," etc. Through semantic comparison, the semantic matching degree between each functionally related fragment and the text fragment extracted from the user's historical conversation text is calculated. When the semantic matching degree reaches a preset requirement (e.g., 0.7 or higher), the functionally related fragment is retained; otherwise, it is excluded. During the comparison process, the semantic similarity of the text and the consistency of business logic need to be considered to ensure that the retained fragments can be effectively semantically associated with the user's requirement statement.

[0052] Step S117: Extract the text fragment containing the user's immediate needs from the current session input.

[0053] In this embodiment, a semantic parsing method similar to step S1132 is used to parse the current session input content and extract text fragments that contain the user's immediate needs. For example, if the current session input is "I want to upgrade my digital service, but I don't know what materials I need to prepare?", semantic parsing extracts the text fragment "What materials are needed to upgrade my digital service?". During the extraction process, it is necessary to accurately identify the user's immediate needs, exclude irrelevant information such as greetings and interjections, and ensure that the extracted text fragments completely express the user's immediate needs.

[0054] Step S118: Semantically correlate the text fragment containing the user's immediate needs extracted from the current session input with the fragments retained in the user's historical session text and the fragments retained in the digital service function description text to supplement the function association information corresponding to the user's immediate needs.

[0055] In this embodiment, a deep learning-based text semantic association model (such as a BERT-based text semantic association model) is used to semantically associate text fragments containing the user's immediate needs extracted from the current session input with fragments retained in the user's historical session text and fragments retained in the digital service function description text. For example, the text fragment extracted from the current session input is "What materials are needed to upgrade a digital service?", the fragment retained in the user's historical session text is "I need to upgrade a digital service, what are the specific steps?", and the fragments retained in the digital service function description text are "Business upgrade function: Log in to the platform to enter the business management module", "Business upgrade function: Select business type and package", "Business upgrade function: Upgrade configuration after payment", "Business upgrade function: View status in my business", etc. Through semantic association, functional association information related to the user's immediate needs is found. Suppose there is a retained fragment in the digital service function description text: "Business upgrade function: The materials needed to upgrade the business include a copy of the ID card, a business application form, and a copy of the original business contract", then this fragment is used as supplementary functional association information and associated with the text fragments pointing to the user's immediate needs. During the association process, it is necessary to ensure that the supplementary functional association information can accurately meet the user's immediate needs, while also effectively semantically connecting with relevant fragments in the user's historical conversation text.

[0056] Step S119: Integrate all retained fragments and fragments supplemented with information related to users' immediate needs, and uniformly label them as conversation interaction gene fragments. Each conversation interaction gene fragment is labeled with its corresponding semantic category and functional association attribute.

[0057] In this embodiment, fragments retained from the user's historical conversation text, fragments retained from the digital service function description text, and fragments supplemented with information related to the user's immediate needs are integrated into a set. Then, each fragment is labeled with a semantic category and functional association attribute. For example, the fragment "I need to upgrade my digital service business, what are the specific steps?" retained from the user's historical conversation text is labeled with the semantic category "User Needs - Business Upgrade Operation Step Consultation" and the functional association attribute "Business Upgrade Function - Operation Step Consultation"; the fragment "Business Upgrade Function: Log in to the platform to enter the business management module" retained from the digital service function description text is labeled with the semantic category "Function Description - Business Upgrade Operation Steps" and the functional association attribute "Business Upgrade Function - Interaction Method Attribute"; the supplemented fragment "Business Upgrade Function: The materials required for business upgrade include a copy of the ID card, a business application form, and a copy of the original business contract" is labeled with the semantic category "Function Description - Business Upgrade Material Requirements" and the functional association attribute "Business Upgrade Function - Core Role Attribute (Material Requirements)". The above annotations ensure that each conversational interaction gene fragment has a clear semantic category and functional association attributes, facilitating subsequent processing and analysis.

[0058] Step S120: Based on the conversation interaction gene fragments, associate the semantic inheritance relationship and functional adaptation relationship between each conversation interaction gene fragment to construct a digital service conversation interaction gene chain. The digital service conversation interaction gene chain includes multiple gene nodes and the association links between gene nodes.

[0059] Step S121: Analyze the semantic category and functional association attributes of each conversation interaction gene segment to determine the core semantic orientation and functional association direction of each conversation interaction gene segment.

[0060] In this embodiment, a semantic analysis model based on the Transformer architecture is used to parse the semantic category and functional association attributes of each conversation interaction gene segment. For example, for the conversation interaction gene segment "I need to upgrade my digital service business, what are the specific steps?", its semantic category is "User Needs - Consultation on Business Upgrade Operation Steps", and its functional association attribute is "Business Upgrade Function - Consultation on Operation Steps". Through parsing, its core semantic focus is determined to be "Consultation on Business Upgrade Operation Steps", and its functional association direction is "Query on the operation steps of the business upgrade function". For the conversation interaction gene segment "Business Upgrade Function: Log in to the platform to enter the business management module", its semantic category is "Function Description - Business Upgrade Operation Steps", and its functional association attribute is "Business Upgrade Function - Interaction Method Attribute". After parsing, its core semantic focus is determined to be "Business Upgrade Operation Steps - Log in to the platform to enter the business management module", and its functional association direction is "Explanation of the interaction method of the business upgrade function". During the parsing process, it is necessary to accurately identify the core semantic focus and functional association direction of each conversation interaction gene segment to ensure that subsequent association analysis can be performed accurately.

[0061] Step S122: Compare the core semantic orientation of any two conversational interaction gene segments to determine whether there is a semantic continuity relationship between the two conversational interaction gene segments. The semantic continuity relationship means that the semantic expression of the previous conversational interaction gene segment provides a premise or supplement to the semantic expression of the next conversational interaction gene segment.

[0062] Step S1221: Extract the core semantic vocabulary set of the first conversational interaction gene segment, wherein the core semantic vocabulary set contains all words representing the core semantics in the conversational interaction gene segment.

[0063] In this embodiment, a rule-based vocabulary extraction method or a deep learning-based vocabulary extraction model is used to extract the core semantic vocabulary of the first conversational interaction gene segment. For example, for the conversational interaction gene segment "I need to upgrade my digital service business, what are the specific steps?", the extracted core semantic vocabulary set is {"business upgrade", "operation steps"}; for the conversational interaction gene segment "business upgrade function: log in to the platform to enter the business management module", the extracted core semantic vocabulary set is {"business upgrade", "log in to the platform", "business management module"}. During the extraction process, it is necessary to ensure that the core semantic vocabulary accurately represents the core semantics of the conversational interaction gene segment and exclude some irrelevant words.

[0064] Step S1222: Extract the core semantic vocabulary set of the second conversational interaction gene segment, compare the core semantic vocabulary set of the first conversational interaction gene segment with the core semantic vocabulary set of the second conversational interaction gene segment, and count the number of identical or synonymous words in the two core semantic vocabulary sets.

[0065] In this embodiment, the same vocabulary extraction method as described above is used to extract the core semantic vocabulary set of the second conversational interaction gene segment, and then the two core semantic vocabulary sets are compared. For example, the core semantic vocabulary set of the first conversational interaction gene segment is {"business upgrade", "operation steps"}, and the core semantic vocabulary set of the second conversational interaction gene segment is {"business upgrade", "login platform", "business management module"}. Through comparison, the number of identical or synonymous words is counted as 1 ("business upgrade" is considered as an identical word). During the comparison process, the semantic similarity of words needs to be considered. For synonymous words (such as "business upgrade" and "service upgrade"), they are also counted as identical or synonymous words.

[0066] Step S1223: Determine whether the number of identical or near-synonyms obtained from the statistics reaches the preset ratio. If the number of identical or near-synonyms reaches the preset ratio, further analyze the semantic expression logic of the first and second conversational interaction gene segments.

[0067] In this embodiment, the preset ratio can be set according to business needs and the accuracy requirements of semantic analysis, for example, set to 0.5. When the number of identical or near-synonyms obtained from the statistics reaches the preset ratio, such as in the example above where the number of identical or near-synonyms is 1 and the total number of words is 2, the ratio is 0.5, reaching the preset ratio, the semantic expression logic of the two conversation interaction gene segments is further analyzed. For example, the semantic expression logic of the first conversation interaction gene segment is the user asking about the operation steps for business upgrade, and the semantic expression logic of the second conversation interaction gene segment is introducing the first operation step of business upgrade (logging into the business management module on the platform). During the analysis process, it is necessary to combine the business scenario and contextual information to accurately understand the semantic expression logic of the two conversation interaction gene segments.

[0068] Step S1224: Determine whether the semantics of the first conversational interaction gene segment provides background information or preconditions for the semantics of the second conversational interaction gene segment.

[0069] In this embodiment, based on the semantic representation logic of the two conversational interaction gene segments, it is determined whether the semantics of the first conversational interaction gene segment provides background information or prerequisites for the semantics of the second conversational interaction gene segment. For example, the semantics of the first conversational interaction gene segment is a user's inquiry about the steps for upgrading a service. This provides background information and prerequisites for the semantics of the second conversational interaction gene segment (introducing the steps for upgrading the service), because without the user's inquiry, there would be no introduction to the steps. If the semantics of the first conversational interaction gene segment provides background information or prerequisites for the semantics of the second conversational interaction gene segment, it is determined that there is a forward semantic continuation relationship between the two conversational interaction gene segments. The forward semantic continuation relationship means that the preceding conversational interaction gene segment provides prerequisites for the following conversational interaction gene segment.

[0070] Step S1225: If the semantics of the first conversational interaction gene segment does not provide background information or preconditions for the semantics of the second conversational interaction gene segment, determine whether the semantics of the second conversational interaction gene segment provides supplementary explanation or extended description for the semantics of the first conversational interaction gene segment.

[0071] In this embodiment, if the semantics of the first session interaction gene segment does not provide background information or preconditions for the semantics of the second session interaction gene segment, it is determined whether the semantics of the second session interaction gene segment provides supplementary explanation or extended description of the semantics of the first session interaction gene segment. For example, if the semantics of the first session interaction gene segment is "Business upgrade function: Log in to the platform to enter the business management module," and the semantics of the second session interaction gene segment is "Business upgrade function: Select business type and package," then the semantics of the second session interaction gene segment is a supplementary explanation and extended description of the semantics of the first session interaction gene segment because it introduces the next step in the business upgrade operation. If the semantics of the second session interaction gene segment provides supplementary explanation or extended description of the semantics of the first session interaction gene segment, it is determined that there is a backward semantic continuation relationship between the two session interaction gene segments. The backward semantic continuation relationship means that the latter session interaction gene segment supplements the former session interaction gene segment.

[0072] Step S1226: If the semantic representations of the two conversational interaction gene fragments do not have obvious premises or complementary relationships, but have a common semantic theme, analyze the order of appearance of the two conversational interaction gene fragments in the user's historical conversation text.

[0073] In this embodiment, if the semantic representations of two conversational interaction gene fragments do not have an obvious premise or complementary relationship, but share a common semantic theme (such as both being related to business upgrade functions), the order in which the two conversational interaction gene fragments appear in the user's historical conversation text is analyzed. For example, if the first conversational interaction gene fragment appears as the 5th occurrence in the user's historical conversation text, and the second conversational interaction gene fragment appears as the 6th occurrence, the semantic continuity relationship between the two conversational interaction gene fragments is determined based on their order of appearance.

[0074] Step S1227: Determine the semantic continuity relationship between the two conversation interaction gene fragments according to the order in which they appear in the user's historical conversation text.

[0075] In this embodiment, based on the order in which two conversational interaction gene fragments appear in the user's historical conversation text, it is determined that the preceding conversational interaction gene fragment is a forward semantic continuation of the following conversational interaction gene fragment, or the following conversational interaction gene fragment is a backward semantic continuation of the preceding conversational interaction gene fragment. For example, if the first conversational interaction gene fragment appears as the 5th instance in the user's historical conversation text, and the second conversational interaction gene fragment appears as the 6th instance, then the first conversational interaction gene fragment is a forward semantic continuation of the second conversational interaction gene fragment, and the second conversational interaction gene fragment is a backward semantic continuation of the first conversational interaction gene fragment.

[0076] Step S1228: If the number of identical or near-synonyms does not reach the preset ratio, and the two conversational interaction gene segments do not share a common semantic topic, it is determined that there is no semantic continuity relationship between the two conversational interaction gene segments.

[0077] In this embodiment, if the number of identical or near-synonymous words does not reach a preset ratio, and the two conversational interaction gene segments do not share a common semantic theme (such as one being related to business upgrade functions and the other to data backup functions), it is determined that there is no semantic continuity relationship between the two conversational interaction gene segments.

[0078] Step S1229: Record all conversational interaction gene fragment pairs with preceding and following semantic continuation relationships and their continuation types, wherein the continuation types include forward semantic continuation relationship types or backward semantic continuation relationship types.

[0079] In this embodiment, all conversational interaction gene fragment pairs with preceding and following semantic continuation relationships are recorded and their continuation types are labeled. For example, the continuation type of the conversational interaction gene fragment pair ("I need to upgrade my digital service business, what are the specific steps?", "Business upgrade function: Log in to the platform to enter the business management module") is a forward semantic continuation relationship; the continuation type of the conversational interaction gene fragment pair ("Business upgrade function: Log in to the platform to enter the business management module", "Business upgrade function: Select business type and package") is a backward semantic continuation relationship. This recording facilitates subsequent association analysis and gene chain construction.

[0080] Step S123: For two conversational interaction gene segments that have a semantic continuity relationship, establish a semantic inheritance association between the two conversational interaction gene segments, and record the semantic continuity direction and continuity strength of the semantic inheritance association.

[0081] In this embodiment, a semantic inheritance association is established for two conversational interaction gene segments that have a semantic continuity relationship. For example, for a conversational interaction gene segment pair ("I need to upgrade my digital service business, what are the specific steps?", "Business upgrade function: Log in to the platform and enter the business management module"), a semantic inheritance association is established, where the semantic continuity direction is from the first conversational interaction gene segment to the second conversational interaction gene segment. The continuity strength can be calculated by factors such as the number of identical or synonymous words and the tightness of the semantic expression logic. For example, the more identical or synonymous words and the tighter the semantic expression logic, the higher the continuity strength. During the recording process, it is necessary to accurately record the semantic continuity direction and continuity strength of the semantic inheritance association to ensure the accurate construction of the gene chain.

[0082] Step S124: Analyze the functional association attributes of two conversational interaction gene segments that have semantic inheritance relationships, and determine whether the functional association direction of the former conversational interaction gene segment matches the functional association direction of the latter conversational interaction gene segment.

[0083] Step S1241: Extract the functional association attributes of the previous conversational interaction gene segment with semantic inheritance association, and determine the functional module type, functional scope and functional interaction requirements corresponding to the conversational interaction gene segment.

[0084] In this embodiment, a rule-based attribute extraction method or a deep learning-based attribute extraction model is used to extract the functional association attributes of the previous conversation interaction fragment. For example, for the previous conversation interaction fragment "I need to upgrade my digital service business, what are the specific steps?", its functional association attribute is "Business Upgrade Function - Operation Step Consultation". Through extraction, the corresponding functional module type is determined to be "Business Upgrade Function Module", the functional scope is "Business Upgrade Operation Step Consultation", and the functional interaction requirement is "User submits consultation, digital server provides operation step instructions". During the extraction process, it is necessary to accurately identify the functional module type, functional scope, and functional interaction requirements to ensure accurate subsequent adaptation judgment.

[0085] Step S1242: Extract the functional association attributes of the next conversational interaction gene segment that has a semantic inheritance relationship, and determine the functional module type, functional scope and functional interaction requirements corresponding to the conversational interaction gene segment.

[0086] In this embodiment, the same attribute extraction method described above is used to extract the functional association attributes of the subsequent session interaction gene segment. For example, for the subsequent session interaction gene segment "Business upgrade function: Log in to the platform to enter the business management module", its functional association attribute is "Business upgrade function - interaction method attribute". After extraction, its corresponding functional module type is determined to be "Business upgrade function module", its functional scope is "Operation steps for business upgrade - Log in to the platform to enter the business management module", and its functional interaction requirement is "User logs in to the platform and enters the business management module". During the extraction process, it is necessary to ensure the accuracy of the extraction of the functional module type, functional scope, and functional interaction requirements.

[0087] Step S1243: Determine whether the functional module corresponding to the previous session interaction gene fragment provides data support or operational basis for the functional module corresponding to the next session interaction gene fragment.

[0088] In this embodiment, based on the functional module type, functional scope, and functional interaction requirements corresponding to the preceding and following session interaction gene segments, it is determined whether the preceding functional module provides data support or operational basis for the following functional module. For example, if the functional scope of the preceding functional module is consultation on the operational steps of business upgrades, and the functional scope of the following functional module is the operational steps of business upgrades—logging into the platform and entering the business management module, the consultation requirement of the preceding functional module provides the operational basis for the operational steps of the following functional module, because users only need to understand the specific operational steps when they make a consultation. If the functional module corresponding to the preceding session interaction gene segment provides data support or operational basis for the functional module corresponding to the following session interaction gene segment, it is determined that there is a basic fit in the functional association direction between the preceding and following session interaction gene segments.

[0089] Step S1244: If the functional module corresponding to the previous session interaction gene segment does not provide data support or operational basis for the functional module corresponding to the next session interaction gene segment, determine whether the functional module corresponding to the next session interaction gene segment is an extension function or a collaborative function of the functional module corresponding to the previous session interaction gene segment.

[0090] In this embodiment, if the preceding functional module does not provide data support or operational basis for the following functional module, it is determined whether the following functional module is an extension or collaborative function of the preceding functional module. For example, if the scope of the preceding functional module is consultation on the operational steps of business upgrades, and the scope of the following functional module is consultation on service usage after the business upgrade, the following functional module is an extension of the preceding functional module because it provides consultation on service usage after the business upgrade is completed. If the functional module corresponding to the following session interaction gene segment is an extension or collaborative function of the functional module corresponding to the preceding session interaction gene segment, it is determined that there is an extended adaptation in the functional association direction between the preceding and following session interaction gene segments.

[0091] Step S1245: If the functional module corresponding to the next session interaction gene segment is not an extension or synergistic function of the functional module corresponding to the previous session interaction gene segment, determine whether the functional module corresponding to the previous session interaction gene segment and the functional module corresponding to the next session interaction gene segment belong to the same functional system.

[0092] In this embodiment, if the subsequent functional module is not an extension or collaborative function of the preceding functional module, it is determined whether the preceding and subsequent functional modules belong to the same functional system. For example, if the preceding functional module is a business upgrade functional module and the subsequent functional module is a data backup functional module, they do not belong to the same functional system; however, if the preceding functional module is a business upgrade functional module and the subsequent functional module is a business downgrade functional module, they belong to the same functional system (business management functional system). During the determination process, it is necessary to analyze the functional architecture of the business to ensure the accuracy of the judgment.

[0093] Step S1246: Determine whether there is a conflict between the functional scope of the functional module corresponding to the previous session interaction gene fragment and the functional module corresponding to the next session interaction gene fragment.

[0094] In this embodiment, if the preceding and following functional modules belong to the same functional system, their functional scope is further analyzed. For example, if the preceding module is a service upgrade functional module (scope: improving service benefits), and the following module is a service downgrade functional module (scope: reducing service benefits), although their scopes are opposite, they serve the overall functional system of service adjustment and there is no conflict. However, if the preceding module is a service upgrade functional module (scope: improving the bandwidth service of a certain service), and the following module is also a service upgrade functional module but its scope is to reduce the storage service of that service, then there is a conflict in the functional scope. If there is no conflict in the functional scope, it is determined that the functional association direction of the preceding and following session interaction gene segments is system-compatible.

[0095] Step S1247: If any of the basic adaptation, extended adaptation, or system adaptation exists, confirm that the functional association direction between the previous session interaction gene segment and the next session interaction gene segment forms an adaptation; if the functional module corresponding to the previous session interaction gene segment and the functional module corresponding to the next session interaction gene segment have no data support relationship, are not extended or collaborative functions, and do not belong to the same functional system, confirm that the functional association direction between the previous session interaction gene segment and the next session interaction gene segment does not form an adaptation.

[0096] In this embodiment, functional association is considered compatible if any one of the following conditions is met: basic, extended, or system adaptation. For example, if the preceding module provides the operational basis for the following module (basic adaptation), or the following module is an extension of the preceding module (extended adaptation), or both belong to the same functional system and their scopes do not conflict (system adaptation), then adaptation is confirmed. Conversely, if none of these relationships exist, then adaptation is confirmed as not occurring.

[0097] Step S1248: Form matching conversational interaction gene fragment pairs for functional association directions, label the adaptation type, the adaptation type includes basic adaptation type, extended adaptation type or system adaptation type, and record the functional association basis for the adaptation.

[0098] In this embodiment, for adapted fragment pairs, their adaptation types are marked. For example, if the preceding module provides the operational basis for the following module, it is marked as a basic adaptation type, and the basis is recorded as: "The consultation needs of the preceding module provide the operational basis for the operation steps of the following module"; if the following module is an extension function of the preceding module, it is marked as an extended adaptation type, and the basis is recorded as: "The service usage consultation of the following module is an extension need after the business upgrade of the preceding module is completed"; if the two belong to the same functional system and their scope of action does not conflict, they are marked as a system adaptation type, and the basis is recorded as: "Both the preceding and following modules belong to the business management functional system, and their functional scopes do not conflict".

[0099] Step S125: Take each session interaction gene fragment as a gene node, connect gene nodes with semantic inheritance association or functional adaptation association through association links, and mark the association type on the association links. The association type includes semantic inheritance association type or functional adaptation association type.

[0100] In this embodiment, each session interaction gene segment is treated as a gene node. For nodes with semantic inheritance or functional adaptation associations, they are connected by association links. For example, if node A (user inquires about business upgrade operation steps) and node B (introduces the first step of business upgrade: logging into the platform) have semantic inheritance and basic adaptation associations, then A and B are connected by a link labeled "semantic inheritance association (forward), basic adaptation". Through the above method, an association network between gene nodes is constructed.

[0101] Step S126: Adjust the order of gene nodes according to the temporal and semantic logical order of the conversation interaction, so that the order of gene nodes conforms to the natural flow of conversation interaction and semantic evolution logic.

[0102] In this embodiment, the arrangement of gene nodes is adjusted by combining the temporal sequence of the conversation (e.g., the user first inquires, then receives instructions on the operation steps, and then inquires about material preparation) and the semantic logical sequence (e.g., the operation steps for business upgrades have a sequential order). For example, the node order should be: user inquiry node → operation step 1 node → operation step 2 node → material preparation inquiry node → material preparation instructions node, to ensure that it conforms to the natural flow of the conversation and the semantic evolution logic.

[0103] Step S127: Identify whether there are isolated gene nodes in the digital service session interaction gene chain, wherein an isolated gene node refers to a gene node that has no associated link connection.

[0104] In this embodiment, all gene nodes are traversed, and each node is checked to see if it has at least one associated link. If a node has no links connected to it, it is an isolated node. For example, if a node describes a data backup function, but no other node is associated with it in the current session, it is identified as an isolated node.

[0105] Step S128: Analyze the semantic and functional attributes of the identified isolated gene nodes.

[0106] In this embodiment, the semantic category and functional association attributes of isolated nodes are analyzed. For example, if the semantic category of an isolated node is "functional description - data backup function" and the functional association attribute is "data backup function - core function attribute", its semantic orientation (introduction of data backup function) and functional association direction (description of the core function of data backup function) are analyzed.

[0107] Step S129: Based on the semantic and functional attributes of isolated gene nodes, supplement the association between isolated gene nodes and other gene nodes, and eliminate isolated gene nodes.

[0108] In this embodiment, based on the semantic and functional attributes of the isolated node, other nodes that are semantically or functionally related to it are found to supplement the association. For example, if the isolated node is a description of the data backup function, and there is a node where users inquire about business data security, then the semantic inheritance relationship between the two is supplemented (users inquiring about data security provides a premise for the description of the data backup function) and the functional adaptation relationship (the data backup function provides protection for business data security, which is a basic adaptation), thereby eliminating the isolated node.

[0109] Step S1210: Integrate all gene nodes, associated links and associated types to form a digital service session interaction gene chain, wherein the digital service session interaction gene chain also contains evolutionary time sequence information corresponding to the gene node arrangement order.

[0110] In this embodiment, all gene nodes, associated links, association types, and the arrangement order of nodes (evolutionary time sequence information) are integrated to form a digital service session interaction gene chain. For example, the structure of the gene chain is: Node 1 (user consultation service upgrade operation steps) → Link (semantic inheritance, basic adaptation) → Node 2 (introduction of the first step of service upgrade: logging into the platform) → Link (semantic inheritance, extended adaptation) → Node 3 (introduction of the second step of service upgrade: selecting a package) → ... → Node n (data backup function description, associated with the user consultation data security node), while recording the time sequence and semantic logical order information of each node.

[0111] Step S130: Extract the user's immediate session intention based on the digital service session interaction gene chain and the user's current session input content, associate the user's immediate session intention with the function association information in the digital service session interaction gene chain, and generate an intention symbiotic evolution unit.

[0112] Step S131: Analyze the user's current session input content, and extract the immediate demand words in the user's current session input content. The immediate demand words refer to the most core words in the user's current expressed demand and can represent the user's immediate appeal.

[0113] Step S1311: Clean the text of the user's current session input content, and split the cleaned user's current session input content into multiple consecutive word sequences.

[0114] In this embodiment, clean the text of the user's current input content, remove stop words (such as "de", "le", etc.) and special characters (such as @, #, etc.), and then split it into word sequences according to natural semantics. For example, if the user inputs "I want to handle the business upgrade of digital services now, but I don't know what materials are needed?", after cleaning, it is split into word sequences such as ["handle", "digital service", "business upgrade", "prepare", "materials"].

[0115] Step S1312: Analyze the semantic weight of each word sequence in the user's current session input content. The semantic weight refers to the importance of the word sequence for expressing the user's immediate demand.

[0116] In this embodiment, adopt a semantic weight calculation method based on word vectors to analyze the semantic weight of each word sequence. For example, the semantic weights of "business upgrade" and "materials" are relatively high because they directly reflect the user's needs; the semantic weights of "handle" and "digital service" are relatively low. When calculating, factors such as the position of the word in the sentence and its词性 (such as noun, verb) can be combined. The more core the position and the more critical the词性, the higher the semantic weight.

[0117] Step S1313: Screen out the word sequence with the highest semantic weight. Among them, if there are multiple word sequences with the same semantic weight, analyze the appearance position of the word sequence in the user's current session input content, and preferentially select the word sequence located at the core position of the user's current session input content.

[0118] In this embodiment, the word sequence with the highest semantic weight is selected. If multiple words have the same weight, such as "business upgrade" and "materials" having the same weight, their positions are analyzed. "Business upgrade" is usually in the middle or early part of the sentence, while "materials" is at the end. Words in core positions (such as the middle) usually better represent the need, so "business upgrade" is selected. Alternatively, if the user's need is to inquire about the preparation of materials for a business upgrade, then both "materials" and "business upgrade" are important, and it may be necessary to make a judgment based on the business scenario. Assuming that "business upgrade" is in a core position based on position, it is selected first; if the positions are close, both can be retained.

[0119] Step S1314: Semantically simplify the selected word sequence, remove the modifying words in the word sequence. If the simplified core words contain multiple semantic references, combine the common demand semantic types in the digital service session interaction gene chain to determine the words corresponding to the semantic references most closely related to the digital service functions, and mark the determined words as the immediate demand words of the user's current session input content.

[0120] In this embodiment, the selected word sequence is simplified. For example, "business upgrade of digital services" is simplified to "business upgrade". Here, it is assumed that "digital services" is a modifier, but it should actually be retained. Alternatively, in a new example: the user inputs "I want to quickly process the business upgrade of your platform. What materials do I need to prepare?", and the selected word sequence is "quickly process", "business upgrade", "prepare", "materials". After semantic simplification, the modifier "quickly" in "quickly process" is removed, resulting in "process", "business upgrade", "prepare", "materials". If the core words after simplification have multiple semantic references, such as "business upgrade" and "materials", combined with the common semantic types of needs in the gene chain (such as the need types related to business upgrades, such as operation steps, material preparation, rights consultation, etc.), "materials" is most closely related to the material preparation needs of business upgrades. Therefore, "materials" and "business upgrade" are determined as immediate need words and marked as the user's immediate need words.

[0121] Step S132: Based on the extracted immediate demand vocabulary, and combined with the core semantic orientation of the gene nodes in the digital service session interaction gene chain, determine the target gene node that semantically matches the user's immediate demand vocabulary.

[0122] In this embodiment, the extracted immediate demand terms (such as "business upgrade" and "materials") are matched with the core semantic references of each node in the gene chain. The core semantic references of the nodes in the gene chain are such as "business upgrade operation steps consultation" and "business upgrade material requirements description". Nodes that semantically match "business upgrade" and "materials" are found. For example, the node whose core semantic reference is "business upgrade material requirements description" is identified as the target gene node.

[0123] Step S133: Extract the functional association information of the target gene node, which includes the functional module description, functional applicable conditions and functional interaction process corresponding to the gene node.

[0124] In this embodiment, functional association information is extracted from the target gene node. For example, the target gene node is "Business upgrade function: The materials required for business upgrade include a copy of the ID card, a business application form, and a copy of the original business contract". Its functional module description is "Business upgrade function", the applicable condition for the function is "when applying for a business upgrade", and the functional interaction process is "the user prepares the materials and submits them".

[0125] Step S134: Integrate the functional association information of immediate need words and target gene nodes to form a complete expression of the user's immediate conversation intent. The complete expression includes the user's immediate need direction and the corresponding functional matching direction.

[0126] In this embodiment, the functional association information of the immediate need terms ("business upgrade" and "materials") and the target gene node is integrated to form a complete statement: "The user's immediate need is to consult the materials required for business upgrade, and the corresponding functional matching direction is the material preparation requirements description of the business upgrade functional module."

[0127] Step S135: Traverse the surrounding gene nodes in the digital service session interaction gene chain that have a related link with the target gene node.

[0128] In this embodiment, the associated links of the target gene node (such as the business upgrade material requirements description node) are found, and the surrounding nodes connected to it are traversed, such as the business upgrade operation steps node, the business upgrade rights description node, etc.

[0129] Step S136: Extract functional association information of surrounding gene nodes and determine the associated functional information that can form a synergy with the user's real-time conversation intent.

[0130] In this embodiment, functional association information of surrounding nodes is extracted. For example, the functional association information of the business upgrade operation steps node is "login to platform, select package, submit materials". The system analyzes whether this information can be coordinated with the user's immediate intention (consulting on material preparation). The information of "submit materials" is coordinated with the intention of material preparation, so the "operation process of submitting materials" is determined to be the associated functional information.

[0131] Step S137: Semantically fuse the complete expression of the user's real-time conversation intent with the determined associated functional information to obtain fused content.

[0132] In this embodiment, the user's immediate intent expression ("materials required for upgrading the consultation service") and related functional information ("the operation process for submitting materials") are combined to obtain the fused content: "materials required for upgrading the user's consultation service, and the operation process for submitting materials".

[0133] Step S138: Divide the semantic levels of the integrated content, taking the user's immediate needs as the core level and related functional information as the extended level, to form a hierarchical intent expression structure.

[0134] In this embodiment, semantic levels are divided: the core level is "materials required for consulting service upgrade" (user's immediate needs), and the extended level is "operation process for submitting materials" (related functional information), forming a structure: core level → extended level.

[0135] Step S139: Based on the hierarchical intent representation structure, generate an intent co-evolution unit. The intent co-evolution unit includes a core intent layer, an extended intent layer, and a semantic association description between the core intent layer and the extended intent layer. The core intent layer refers to the representation content corresponding to the user's immediate needs, and the extended intent layer refers to the representation content corresponding to the associated functional information.

[0136] In this embodiment, the intention co-evolution unit is generated as follows: the core intention layer is "materials required for consulting business upgrade", the extended intention layer is "operation process for submitting materials", and the semantic association description is "the operation process for submitting materials is a follow-up step after the preparation of business upgrade materials is completed, and it is coordinated with the material preparation requirements".

[0137] Step S150: Receive feedback text from the user regarding the intention co-evolution unit, analyze the semantic deviation between the feedback text and the intention co-evolution unit, and correct the evolution path of the digital service session interaction gene chain based on the semantic deviation to obtain the corrected evolution path.

[0138] In this embodiment, user feedback text is received through an input interface that interacts with the user. This interface supports input formats such as text and voice, and the input voice content is converted into text format in real time. For example, after receiving the intention symbiotic evolution unit (the core intention layer is "materials required for consulting business upgrades," the extended intention layer is "the operation process for submitting materials," and the semantic association is "the operation process for submitting materials is the subsequent step after the preparation of business upgrade materials is completed, and it coordinates with the material preparation requirements"), the user inputs the feedback text "I also want to know the review time after the materials are submitted and the handling method if the review fails." The feedback text is semantically parsed to extract core semantic words, such as "material submission," "review time," "review failure," and "handling method." Then, the semantic content of the intention symbiotic evolution unit is analyzed. Its core intention is consulting business upgrade materials, its extended intention is the material submission process, and its semantic association is the coordination between the submission process and material preparation. By comparing the core semantics of the feedback text with the semantics of the intent co-evolution unit, it was found that the semantic deviation of the feedback text lies in the expansion of the scope of the requirement, from material preparation and submission process to review time and handling methods for failed reviews. This is a newly added dimension of requirement, and although this requirement is related to the material submission stage of the business upgrade, it was not reflected in the original intent co-evolution unit. Based on this semantic deviation, the existing evolution path of the digital service session interaction gene chain was analyzed. The original path included nodes such as the business upgrade material requirement description node and the material submission operation process node. To correct the evolution path, nodes related to review time and handling methods for failed reviews need to be added to the gene chain, such as the "Business Upgrade Material Review Time Description" node and the "Business Upgrade Material Failure Handling Method Description" node. Furthermore, a link needs to be established between these new nodes and the original "Material Submission Operation Process" node. The type of link is clearly defined as semantic inheritance link (because review time and handling methods are subsequent stages after material submission, the semantics of the original submission process node provides a premise for the new nodes) and functional adaptation link (both the review stage and the material submission stage belong to the business upgrade process system, and their functional scopes do not conflict, thus belonging to system adaptation). The above corrections yield a revised evolution path that includes new nodes and associated links.

[0139] Step S160: According to the modified evolution path, update the gene node association relationship and the content of the intention co-evolution unit in the digital service session interaction gene chain to obtain the iterated digital service session interaction gene chain.

[0140] In this embodiment, following the revised evolution path, the association relationships of gene nodes are first updated. For the newly added "Business Upgrade Material Review Time Description" and "Business Upgrade Material Failure Handling Method Description" nodes, an association link is established between them and the "Material Submission Operation Process" node. The association type is marked as semantic inheritance association (forward) and system adaptation type, and the adaptation basis is recorded (the review stage is a process stage after material submission, and the two belong to the business upgrade process system, with no conflict in their functional scope). Then, the content of the intent symbiotic evolution unit is updated. The core intent layer retains "Materials required for consulting business upgrades," and the extended intent layer adds "Review Time after Material Submission" and "Handling Method for Failure of Review." The semantic association description is updated to "The material submission operation process is the subsequent step after the business upgrade material preparation is completed, the review time after material submission is the subsequent stage of the submission process, and the handling method for failure of review is an extension of the review stage. The three, together with the material preparation requirements, constitute the relevant requirements of the business upgrade process." The updated gene node relationships and intent symbiotic evolution unit content are integrated into the original digital service session interaction gene chain to obtain the iterative digital service session interaction gene chain. This gene chain contains newly added nodes, updated relationships, and intent symbiotic evolution unit content, which can more comprehensively reflect the user's session needs and the functional relationships of digital services.

[0141] Step S160: Based on the iterative digital service session interaction gene chain, summarize the evolution law of the digital service session interaction gene chain and the intent co-existence adaptation mode, and output the digital service session interaction standard evolution mechanism. The digital service session interaction standard evolution mechanism is used to guide the construction of the digital service session interaction gene chain and the intent co-existence process of subsequent session interactions.

[0142] For example, step S161: compare the digital service session interaction gene chain before and after the iteration, and analyze the change type of gene node association in the digital service session interaction gene chain. The change type includes the addition of association links, the strengthening of association links, the adjustment of association links and the supplementation of gene nodes.

[0143] In this embodiment, the gene chains before and after the iteration are compared to analyze the types of changes. For example, after the iteration, a new link was added between the "Business Upgrade Material Review Time Description" node and the original node (new link); the original link between "Business Upgrade Material Requirements Description" and "Material Submission Operation Process" became closer due to user feedback (link strengthened); the direction of a certain link was adjusted (link adjusted); and new gene nodes were added (gene node supplemented).

[0144] Step S162: Statistically analyze the triggering conditions corresponding to each type of change. The triggering conditions include the semantic deviation type of user feedback, the degree of matching between immediate needs and the digital service session interaction gene chain, and the adaptation of functional association information.

[0145] In this embodiment, the triggering conditions for statistical change types are defined. For example, the triggering condition for adding a new link is that the semantic deviation type reported by the user is "requirement expansion," the matching degree between the immediate requirement and the gene chain is "partial match" (the original gene chain had no review time-related nodes), and the adaptation of the functional association information is that the review time is compatible with the material preparation and submission process. The triggering condition for strengthening the link is that the semantic deviation type reported by the user is "requirement deepening" (inquiries about the details of the submission process), the matching degree between the immediate requirement and the gene chain is "high match," and the adaptation of the functional association information is that the details of the submission process are compatible with the original intent.

[0146] Step S163: Based on the correspondence between change type and triggering condition, summarize the evolution law of digital service session interaction gene chain. The evolution law includes the common change direction and change magnitude of digital service session interaction gene chain under different triggering conditions.

[0147] In this embodiment, the evolutionary pattern is summarized as follows: when the triggering conditions are demand expansion, partial matching, and functional adaptation, the direction of change in the gene chain is to add new related links and nodes, and the magnitude of change is a small-scale expansion; when the triggering conditions are demand deepening, high matching, and functional adaptation, the direction of change is to strengthen related links, and the magnitude of change is an increase in link strength.

[0148] Step S164: Analyze the generation method of intention co-evolution units during multiple iterations, and summarize the intention co-evolution adaptation mode. The intention co-evolution adaptation mode includes the association method between core intention and extended intention, the fusion method between intention and functional information, and the adjustment method of intention as the digital service session interaction gene chain evolves.

[0149] In this embodiment, the generation method of intention co-evolution unit through multiple iterations is analyzed, and the pattern is summarized as follows: the association between core intention and extended intention is mostly demand-subsequent steps or demand-related functions; the fusion method of intention and functional information is semantic integration, retaining core demand and expanding related functional information; the adjustment method of intention with gene chain evolution is to update the content of core intention or extended intention according to the addition of nodes and changes in the link of the gene chain.

[0150] Step S165: Based on the evolutionary rules of the digital service session interaction gene chain, formulate the initial construction process of the digital service session interaction gene chain in subsequent session interactions. The initial construction process includes the priority order of session interaction gene fragment extraction, the judgment criteria for gene node association, and the arrangement principle of the initial structure of the digital service session interaction gene chain.

[0151] In this embodiment, the initial construction process is defined as follows: the priority order for extracting session interaction gene fragments is that fragments related to user needs take precedence over fragments describing functions; the criterion for determining the association of gene nodes is the dual satisfaction of semantic inheritance and functional adaptation; the principle for arranging the initial structure is a combination of chronological order and semantic logical order.

[0152] Step S166: Based on the intent co-existence adaptation mode, formulate the generation process of intent co-existence evolution units in subsequent conversation interactions. The generation process includes an instant intent extraction method, a screening standard for associated functional information, and an intent hierarchy division method.

[0153] In this embodiment, the generation process is defined as follows: the instant intent extraction method adopts the steps of text cleaning, semantic weight analysis, and vocabulary simplification; the screening criteria for related functional information are functional information that is semantically coordinated with the instant intent and belongs to the peripheral nodes of the gene chain; the intent hierarchy division method is that the core intent is the user's instant needs, and the extended intent is related functional information.

[0154] Step S167: Integrate the initial construction process of the digital service session interaction gene chain, the generation process of the intent symbiotic evolution unit, the evolution law of the digital service session interaction gene chain, and the intent symbiotic adaptation mode to form the core content of the digital service session interaction standard evolution mechanism.

[0155] In this embodiment, the above processes, rules, and patterns are integrated to form core content, and the norms for gene chain construction, intent generation, and evolutionary adjustment are clarified.

[0156] Step S168: Supplement the application scope description of the digital service session interaction standard evolution mechanism, and determine the types of digital services, session interaction scenarios and user needs to which the digital service session interaction standard evolution mechanism is applicable.

[0157] In this embodiment, the scope of application is expanded to include various digital services (such as e-commerce, finance, government affairs, etc.), conversational interaction scenarios including consultation, processing, complaints, etc., and user demand types including function consultation, business processing, rights and interests claims, etc.

[0158] Step S169: Determine the update triggering conditions for the evolution mechanism of digital service session interaction standards. The update triggering conditions include digital service function upgrades, significant changes in user session habits, and adjustments to industry service standards.

[0159] In this embodiment, the update trigger conditions are determined as follows: when digital service functions are upgraded (such as adding new business types), user conversation habits change significantly (such as changing the consultation method from text to voice), or industry service standards are adjusted (such as upgrading data security requirements), the standard evolution mechanism is updated.

[0160] Step S1610: Integrate the core content, application scope description and update triggering conditions of the digital service session interaction standard evolution mechanism to form the digital service session interaction standard evolution mechanism.

[0161] In this embodiment, the core content, application scope, and update triggering conditions are integrated to form a complete digital service session interaction standard evolution mechanism, which is used to guide subsequent session interaction processing.

[0162] Based on the same inventive concept, please refer to Figure 2 This paper shows a schematic block diagram of a digital service session interaction processing system 100 provided in an embodiment of the present application for performing the above-described digital service session interaction processing method. The digital service session interaction processing system 100 may include a communication unit 110, a machine-readable storage medium 120, and a processor 130.

[0163] The machine-readable storage medium 120 is used to store machine-executable instructions for executing the scheme of this application, and the processor 130 is used to execute the machine-executable instructions stored in the machine-readable storage medium 120 to implement the digital service session interaction processing method provided in the aforementioned method embodiments.

[0164] It should be noted that, in order to simplify the description of the present invention and thus help to understand one or more embodiments of the invention, multiple features may sometimes be grouped into one embodiment, drawing or description thereof in the foregoing description of the embodiments of the present invention.

Claims

1. A method for processing digital service session interactions, characterized in that, The method includes: Obtain digital service session source data, and extract session interaction gene fragments from the digital service session source data. The digital service session source data includes user historical session text, current session input content, and digital service function description text. The session interaction gene fragments refer to text units that carry the core semantics and functional associations of the session. Based on the aforementioned conversational interaction gene fragments, the semantic inheritance relationship and functional adaptation relationship between each conversational interaction gene fragment are associated to construct a digital service conversational interaction gene chain, which includes multiple gene nodes and the associated links between gene nodes. Based on the digital service session interaction gene chain and the user's current session input, the user's real-time session intent is extracted, and the user's real-time session intent is associated with the functional association information in the digital service session interaction gene chain to generate an intent co-evolution unit. Receive user feedback text for the intent co-evolution unit, analyze the semantic deviation between the feedback text and the intent co-evolution unit, and correct the evolution path of the digital service session interaction gene chain based on the semantic deviation to obtain the corrected evolution path. According to the modified evolution path, update the gene node association relationship and the content of the intention co-evolution unit in the digital service session interaction gene chain to obtain the iterative digital service session interaction gene chain. Based on the iterative digital service session interaction gene chain, the evolution law and intent symbiosis adaptation mode of the digital service session interaction gene chain are summarized, and a standard evolution mechanism for digital service session interaction is output. The standard evolution mechanism for digital service session interaction is used to guide the construction of the digital service session interaction gene chain and the intent symbiosis process of subsequent session interactions.

2. The digital service session interaction processing method according to claim 1, characterized in that, The step of acquiring digital service session source data and extracting session interaction gene fragments from the digital service session source data includes: The user's historical conversation text, current conversation input content, and digital service function description text are separated from the digital service conversation source data, and the data source partitions corresponding to the user's historical conversation text, current conversation input content, and digital service function description text are divided. Cross-compare the text content of the digital service session source data in each partition and remove duplicate text content. The system scans the user's historical conversation text segment by segment to identify paragraphs containing the user's core needs and digital service response semantics. The user's core needs refer to the core demands the user expresses for digital services in the conversation, and the digital service response semantics refer to the functional related responses given by the digital service in response to the user's needs. Extract text segments from the identified paragraphs that can independently carry a single semantic and functional association. The single semantic and functional association refers to the text segment containing only one user demand and the corresponding digital service function association information. Scan the digital service function description text and extract the text content describing the core function, applicable scenarios, and interaction methods of the function. The text content describing the core function, applicable scenarios, and interaction methods of the digital service function will be extracted and broken down into function-related segments that can be associated with the user's needs. The text fragments extracted from the user's historical conversation text are semantically compared with the function-related fragments extracted from the digital service function description text. Fragments with semantic matching degree that meet the preset requirements are retained, while semantically unrelated fragments are excluded. Extract the text fragment containing the user's immediate needs from the current session input; The text fragments containing the user's immediate needs extracted from the current session input are semantically correlated with fragments retained in the user's historical session text and fragments retained in the digital service function description text, supplementing the function association information corresponding to the user's immediate needs. All retained fragments and fragments supplemented with information related to users' immediate needs are integrated and uniformly labeled as conversation interaction gene fragments. Each conversation interaction gene fragment is labeled with its corresponding semantic category and functional association attributes.

3. The digital service session interaction processing method according to claim 2, characterized in that, The step-by-step scanning of user history conversation text to identify paragraphs containing core user needs and digital service response semantics includes: Based on the chronological order of the conversation, the user's historical conversation text is divided into multiple consecutive conversation segments, each containing a complete set of user input and digital service responses. Semantic parsing is performed on each conversation segment to extract the core words of the user input statement in the conversation segment. The core words refer to nouns, verbs or adjectives that can represent the core demands of the user input statement. Based on the extracted core words, determine whether the user input statement contains a definite demand, and mark the conversation paragraph containing the user input statement containing the definite demand as the demand-related paragraph. Parse the digital service response statements in the requirement-related paragraphs, and extract the functional references in the digital service response statements that are related to the user's requirements. The functional references refer to the descriptions of functional modules in the digital service response that can meet the user's requirements. Determine whether the functional references in the digital service response statement semantically correspond to the user's input statement's requirements, and confirm that the semantically corresponding conversation paragraphs contain digital service response semantics; Summarize all conversation paragraphs containing core user needs statements and digital service response semantics to obtain paragraphs containing user needs statements and corresponding digital service response semantics.

4. The digital service session interaction processing method according to claim 2, characterized in that, The process involves scanning the digital service function description text, extracting text content describing the core function, applicable scenarios, and interaction methods, and then breaking down this extracted text content into function-related segments that can be associated with the user's needs. Analyze the digital service function description text sentence by sentence to identify sentences containing function names; Based on the identified function names, the digital service function description text is divided into functional partitions, with each functional partition corresponding to a digital service function module. Within each functional area, extract the text content describing the core function of that functional module. The core function refers to the core problem that the functional module can solve for the user or the core value it can provide. Within each functional partition, extract the text content describing the applicable scenarios of the functional module. The applicable scenarios refer to the conditions under which the functional module can function normally and meet the user's needs. Within each functional area, extract the text content describing the interaction method of that functional module. The functional interaction method refers to the description of the input form, operation steps, and response form required when the user operates the functional module. The text content extracted from each functional area, describing the core function, applicable scenarios, and interaction methods, is integrated into a functional description unit. Each functional description unit corresponds to the complete association information of a functional module. Analyze the semantic length and expression style of user needs in the user's historical conversation text. Based on the semantic length and expression style of user needs in the user's historical conversation text, the functional description unit is divided into multiple short text fragments. The semantic length of each short text fragment is similar to the semantic length of user needs in the user's historical conversation text. Semantic annotation is performed on the split short text fragments, and the functional attributes corresponding to each short text fragment are labeled. The functional attributes include core function attributes, applicable scenario attributes, or interaction method attributes, to obtain functionally associated fragments that can be related to the user's needs.

5. The digital service session interaction processing method according to claim 1, characterized in that, The construction of a digital service conversation interaction gene chain based on the conversation interaction gene fragments, associating the semantic inheritance and functional adaptation relationships between each conversation interaction gene fragment, includes: Analyze the semantic category and functional association attributes of each conversation interaction gene segment to determine the core semantic orientation and functional association direction of each conversation interaction gene segment; By comparing the core semantic orientation of any two conversational interaction gene segments, it is determined whether there is a semantic continuity relationship between the two conversational interaction gene segments. The semantic continuity relationship means that the semantic expression of the previous conversational interaction gene segment provides a premise or supplement to the semantic expression of the subsequent conversational interaction gene segment. For two conversational interaction gene segments that have a semantic continuity relationship, establish a semantic inheritance association between the two conversational interaction gene segments and record the semantic continuity direction and continuity strength of the semantic inheritance association; Analyze the functional association attributes of two conversational interaction gene segments that have semantic inheritance relationships, and determine whether the functional association direction of the former conversational interaction gene segment matches the functional association direction of the latter conversational interaction gene segment. Establish a functional adaptation association between two conversational interaction gene segments that are adapted to each other in terms of functional association direction. Each conversational interaction gene fragment is used as a gene node, and gene nodes with semantic inheritance association or functional adaptation association are connected through association links. The association type is marked on the association link, and the association type includes semantic inheritance association type or functional adaptation association type. The order of gene nodes is adjusted according to the temporal and semantic logical order of the conversation interaction, so that the order of gene nodes conforms to the natural flow of conversation interaction and semantic evolution logic. Identify whether there are isolated gene nodes in the digital service session interaction gene chain, wherein an isolated gene node is a gene node that has no associated link connection. For the identified isolated gene nodes, analyze their semantic and functional attributes; Based on the semantic and functional attributes of isolated gene nodes, the association between isolated gene nodes and other gene nodes is supplemented, and isolated gene nodes are eliminated. By integrating all gene nodes, related links, and related types, a digital service session interaction gene chain is formed, which also includes evolutionary time sequence information corresponding to the gene node arrangement order.

6. The digital service session interaction processing method according to claim 5, characterized in that, The step of comparing the core semantic orientations of any two conversational interaction gene segments to determine whether there is a semantic continuity relationship between the two conversational interaction gene segments includes: Extract the core semantic vocabulary set of the first conversational interaction gene segment, wherein the core semantic vocabulary set contains all words representing the core semantics in the conversational interaction gene segment; Extract the core semantic vocabulary set of the second conversational interaction gene segment, compare the core semantic vocabulary set of the first conversational interaction gene segment with the core semantic vocabulary set of the second conversational interaction gene segment, and count the number of identical or synonymous words in the two core semantic vocabulary sets. Determine whether the number of identical or near-synonyms obtained from the statistics reaches the preset ratio. If the number of identical or near-synonyms reaches the preset ratio, further analyze the semantic expression logic of the first and second conversational interaction gene segments. Determine whether the semantics of the first conversational interaction gene segment provides background information or preconditions for the semantics of the second conversational interaction gene segment; If the semantics of the first conversational interaction gene segment provides background information or preconditions for the semantics of the second conversational interaction gene segment, it is determined that there is a forward semantic continuation relationship between the two conversational interaction gene segments. The forward semantic continuation relationship means that the previous conversational interaction gene segment provides a precondition for the next conversational interaction gene segment. If the semantics of the first conversational interaction gene segment does not provide background information or preconditions for the semantics of the second conversational interaction gene segment, determine whether the semantics of the second conversational interaction gene segment provides supplementary explanation or extended expression for the semantics of the first conversational interaction gene segment. If the semantics of the second conversational interaction gene segment provides supplementary explanation or extended expression to the semantics of the first conversational interaction gene segment, it is determined that there is a backward semantic continuation relationship between the two conversational interaction gene segments. The backward semantic continuation relationship means that the latter conversational interaction gene segment provides supplement to the former conversational interaction gene segment. If the semantic representations of two conversational interaction gene segments do not have obvious premises or complementary relationships, but they share a common semantic theme, analyze the order in which the two conversational interaction gene segments appear in the user's historical conversation text. Determine the semantic continuity relationship between two conversation interaction gene segments based on their order of appearance in the user's historical conversation text. If the number of identical or near-synonymous words does not reach the preset ratio, and the two conversational interaction gene segments do not have a common semantic topic, it is determined that there is no semantic continuity relationship between the two conversational interaction gene segments. Record all conversational interaction gene fragment pairs that have a forward semantic continuation relationship and their continuation types, including forward semantic continuation relationship types or backward semantic continuation relationship types.

7. The digital service session interaction processing method according to claim 5, characterized in that, The analysis of the functional association attributes of two conversational interaction gene segments with semantic inheritance relationships determines whether the functional association direction of the preceding conversational interaction gene segment matches the functional association direction of the following conversational interaction gene segment, including: Extract the functional association attributes of the previous conversational interaction gene segment that has a semantic inheritance relationship, and determine the functional module type, functional scope and functional interaction requirements corresponding to the conversational interaction gene segment; Extract the functional association attributes of the next conversational interaction gene segment that has a semantic inheritance relationship, and determine the functional module type, functional scope and functional interaction requirements corresponding to the conversational interaction gene segment; Determine whether the functional module corresponding to the previous session interaction gene segment provides data support or operational basis for the functional module corresponding to the next session interaction gene segment; If the functional module corresponding to the previous session interaction gene segment provides data support or operational basis for the functional module corresponding to the next session interaction gene segment, it is determined that there is a basic fit between the functional association direction of the previous session interaction gene segment and the next session interaction gene segment. If the functional module corresponding to the previous session interaction gene segment does not provide data support or operational basis for the functional module corresponding to the next session interaction gene segment, determine whether the functional module corresponding to the next session interaction gene segment is an extension or collaborative function of the functional module corresponding to the previous session interaction gene segment. If the functional module corresponding to the next session interaction gene segment is an extension or synergistic function of the functional module corresponding to the previous session interaction gene segment, it is determined that there is an extended adaptation in the functional association direction between the previous session interaction gene segment and the next session interaction gene segment. If the functional module corresponding to the next session interaction gene segment is not an extension or synergistic function of the functional module corresponding to the previous session interaction gene segment, determine whether the functional module corresponding to the previous session interaction gene segment and the functional module corresponding to the next session interaction gene segment belong to the same functional system. Determine whether there is a conflict in the functional scope of the functional module corresponding to the previous session interaction gene segment and the functional module corresponding to the next session interaction gene segment; If the functional module corresponding to the previous session interaction gene segment and the functional module corresponding to the next session interaction gene segment belong to the same functional system and there is no conflict in the scope of function, it is determined that the functional association direction between the previous session interaction gene segment and the next session interaction gene segment has system adaptation. If any of the basic adaptation, extended adaptation, or system adaptation exists, confirm that the functional association direction between the previous session interaction gene segment and the next session interaction gene segment forms an adaptation. If the functional module corresponding to the previous session interaction gene segment has no data support relationship, is not an extended or collaborative function, and does not belong to the same functional system, it is confirmed that the functional association direction between the previous session interaction gene segment and the next session interaction gene segment has not been adapted. For functional association directions, form adapted conversational interaction gene fragment pairs, label the adaptation type, which includes basic adaptation type, extended adaptation type or system adaptation type, and record the functional association basis of the adaptation.

8. The digital service session interaction processing method according to claim 1, characterized in that, The step of extracting the user's real-time conversation intent based on the digital service conversation interaction gene chain and the user's current conversation input content, associating the user's real-time conversation intent with the functional association information in the digital service conversation interaction gene chain, and generating an intent co-evolution unit includes: Parse the user's current conversation input and extract the immediate need words from the user's current conversation input. The immediate need words refer to the core words in the user's current expressed needs, which can represent the user's immediate demands. Based on the extracted immediate demand vocabulary, and combined with the core semantic orientation of the gene nodes in the digital service conversation interaction gene chain, the target gene nodes that semantically match the user's immediate demand vocabulary are determined. Extract the functional association information of the target gene node, which includes the functional module description, functional applicable conditions and functional interaction process of the gene node; Integrate the functional association information of immediate demand vocabulary and target gene nodes to form a complete expression of the user's immediate conversation intent. The complete expression includes the user's immediate demand and the corresponding functional matching direction. Traverse the surrounding gene nodes in the digital service session interaction gene chain that have a link with the target gene node; Extract functional association information from surrounding gene nodes to determine the associated functional information that can form a synergy with the user's real-time conversational intent; The complete expression of the user's real-time conversation intent is semantically fused with the identified related functional information to obtain fused content; The semantic hierarchy of the integrated content is divided, with the user's immediate needs as the core level and related functional information as the extended level, forming a hierarchical intent expression structure. Based on a hierarchical intent representation structure, an intent co-evolution unit is generated. The intent co-evolution unit includes a core intent layer, an extended intent layer, and a semantic association description between the core intent layer and the extended intent layer. The core intent layer refers to the representation content corresponding to the user's immediate needs, and the extended intent layer refers to the representation content corresponding to the associated functional information.

9. The digital service session interaction processing method according to claim 8, characterized in that, The process involves parsing the user's current conversation input and extracting immediate need words from it. These immediate need words refer to the core words representing the user's immediate demands, including: The text is cleaned from the user's current conversation input, and the cleaned user's current conversation input is broken down into multiple consecutive word sequences; Analyze the semantic weight of each word sequence in the user's current session input, where the semantic weight refers to the importance of the word sequence in expressing the user's immediate needs; The word sequence with the highest semantic weight is selected. If there are multiple word sequences with the same semantic weight, the position of the word sequence in the user's current conversation input is analyzed, and the word sequence located in the core position of the user's current conversation input is selected first. The selected word sequence is semantically simplified by removing modifier words. If the simplified core words contain multiple semantic references, the words corresponding to the semantic references most closely related to the digital service functions are determined by combining the common semantic types of demand in the digital service session interaction gene chain. The determined words are marked as the immediate demand words of the user's current session input content.

10. A digital service session interaction processing system, characterized in that, include: processor; A machine-readable storage medium for storing machine-executable instructions of the processor; The processor is configured to execute the digital service session interaction processing method according to any one of claims 1 to 8 by executing the machine-executable instructions.

Images