Conversational processing method, apparatus, device, and program product
By structurally storing users' historical dialogues and generating response content using a large language model, the response problem of intelligent dialogue systems after multiple rounds of dialogue and long periods of interruption has been solved, achieving efficient and accurate dialogue and business processing.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- ANHUI IFLYTEK INTELLIGENT SYST
- Filing Date
- 2026-02-04
- Publication Date
- 2026-06-12
AI Technical Summary
Existing intelligent dialogue systems are unable to respond quickly and accurately to user needs after multiple rounds of dialogue or long periods of interruption, resulting in low dialogue efficiency and business processing efficiency.
By structuring and storing users' historical dialogue content, business memory information and background knowledge information are generated, and response content is generated using a large language model, enabling continuous storage and efficient retrieval across sessions.
It improves the accuracy and efficiency of dialogue, solves the continuity problem across sessions, avoids the waste of storing redundant information, and improves the response speed and business processing efficiency of the dialogue system.
Smart Images
Figure CN122196113A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of artificial intelligence technology, and in particular to a dialogue processing method, apparatus, device, and program product. Background Technology
[0002] With the development of artificial intelligence technology, service scenarios such as intelligent business processing and intelligent customer service are becoming increasingly common. In these business scenarios, intelligent dialogue systems are needed to automatically engage in dialogue with users, thereby assisting them in processing business or responding to their inquiries.
[0003] Current conventional intelligent dialogue systems can accurately achieve real-time "question-and-answer" dialogues. However, when users engage in multiple rounds of dialogue, conduct business multiple times, or resume business operations after a prolonged hiatus, the dialogue system often needs to re-establish the dialogue with the user and repeat past conversations to understand the user's intent or meet their needs. This significantly impacts dialogue and business processing efficiency, hindering the ability to quickly and accurately respond to user dialogues and assist users in completing their business. Summary of the Invention
[0004] In view of the above-mentioned technical problems, this application provides a dialogue processing method, apparatus, device and program product, which can improve the accuracy of responses to business dialogues with users, improve dialogue efficiency and improve the efficiency of users in handling business.
[0005] The first aspect of this application provides a dialogue processing method, including: Obtain the current dialogue content input by the user, wherein the current dialogue content includes content related to handling the target business; Retrieve historical business records from the business record database that correspond to the user and are related to the target business; the historical business records include business memory information and background knowledge information, the business memory information includes the user's historical records of handling the target business, and the background knowledge information includes the user's profile information and / or the business policy information of the target business; If historical business records corresponding to the user and related to the target business are retrieved, a response is generated based on the historical business records and the current conversation content.
[0006] In some implementations, the user's historical records of handling the target service include a summary of the user's historical conversations regarding the target service, as well as the progress and / or status of the user's handling of the target service; The user's profile information is extracted based on the user's historical conversations when handling the target service.
[0007] In some implementations, generating response content corresponding to the current conversation content based on the historical business records and the current conversation content includes: Using the historical business records and the current dialogue content, generate dialogue prompts; The dialogue prompts are input into a large language model, which then generates a response based on the historical business records and the current dialogue content.
[0008] In some implementations, the method further includes: Extract historical business records that correspond to the user and are related to the target business from the user's historical dialogue content when the user handles the target business; The historical business records are stored in the business record database.
[0009] In some implementations, historical business records corresponding to the user and related to the target business are extracted from the user's historical dialogue content when handling the target business, including: The process involves summarizing the historical dialogue content of the user's handling of the target service to obtain a summary of the historical dialogue, determining the progress and / or status of the user's handling of the target service based on the historical dialogue content, and extracting the user's profile information based on the historical dialogue content. The summary of the historical dialogue and the progress and / or status of the user's handling of the target business are stored as business memory information in the historical business record. The user's profile information and the business policy information of the target business are stored as background knowledge information in the historical business record.
[0010] In some implementations, the method further includes: Based on the current dialogue content and / or the response content, update the business memory information and / or the background knowledge information in the historical business records.
[0011] In some implementations, updating the business memory information and / or background knowledge information in the historical business records based on the current dialogue content and / or the response content includes: Based on the current dialogue content and / or the reply content, extract the dialogue summary and the user's profile information, and based on the current dialogue content and / or the reply content, determine the latest progress and / or the latest status of the user's handling of the target business; The extracted dialogue summary is stored in the business memory information in the historical business record, and the progress and / or status of the user's handling of the target business stored in the business memory information is updated using the latest progress and / or status of the user's handling of the target business. If the extracted user profile information and / or the target business policy information are updated relative to the user profile information and / or business policy information stored in the background knowledge information of the historical business records, the extracted user profile information and / or the updated business policy information of the target business shall be used to update the user profile information and / or business policy information stored in the background knowledge information of the historical business records.
[0012] A second aspect of this application provides a dialogue processing apparatus, comprising: The dialogue acquisition unit is used to acquire the current dialogue content input by the user, which includes content related to handling the target business. The data retrieval unit is used to retrieve historical business records corresponding to the user and related to the target business from the business record database; the historical business records include business memory information and background knowledge information, the business memory information includes the user's historical records of handling the target business, and the background knowledge information includes the user's profile information and / or the business policy information of the target business; The dialogue response unit is used to generate response content corresponding to the current dialogue content based on the historical business records that correspond to the user and are related to the target business, when historical business records are retrieved.
[0013] A third aspect of this application provides an electronic device, including a memory and a processor; The memory is connected to the processor and is used to store programs; The processor is used to implement the above-described dialogue processing method by running the program in the memory.
[0014] A fourth aspect of this application provides a computer program product, including computer program instructions, which, when executed by a processor, cause the processor to perform the aforementioned dialogue processing method.
[0015] The dialogue processing method provided in this application extracts business memory information and background knowledge information from the user's historical dialogues related to the target business and stores them as historical business records. When the current dialogue content related to the target business is obtained from the user's input, the method retrieves the historical business records corresponding to the user and related to the target business from the business record database, that is, it retrieves the user's business memory information and background knowledge information related to the target business. Then, based on the retrieved historical business records and the current dialogue content, the method generates a response content corresponding to the current dialogue content.
[0016] The aforementioned method extracts and stores the user's historical dialogues for the target business into business memory information and background knowledge information, achieving structured storage of historical dialogue content. This means it enables persistent and continuous storage of user's historical dialogues across sessions, overcoming the limitations of session cycles and allowing for long-term preservation of business status and user information. Furthermore, by dividing user historical dialogues into business memory information and background knowledge information for structured storage, this application makes the retrieval and updating of historical dialogue content more efficient and accurate, avoiding the waste of storing redundant information and preventing redundant information from negatively impacting subsequent dialogue responses.
[0017] Based on the above characteristics, this application uses the structured historical business information to respond to dialogues, which can improve the accuracy of responses, improve dialogue efficiency, and improve the efficiency of users in handling business. Attached Figure Description
[0018] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of this application. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.
[0019] Figure 1 This is a flowchart illustrating a dialogue processing method provided in an embodiment of this application.
[0020] Figure 2 This is a schematic diagram of the structure of a dialogue processing device provided in an embodiment of this application.
[0021] Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation
[0022] The technical solutions of this application are applicable to application scenarios where intelligent dialogue systems engage in dialogue with users in service scenarios such as intelligent business processing and intelligent customer service. Applying the technical solutions provided in this application to the aforementioned intelligent dialogue system, enabling the system to engage in dialogue with users according to the solutions of this application, can improve the accuracy of dialogue responses, increase dialogue efficiency, and enhance the efficiency of users in processing business.
[0023] The aforementioned intelligent dialogue system can be a dialogue system based on a large language model (LLM).
[0024] Currently, in service scenarios such as intelligent business processing and intelligent customer service, intelligent dialogue systems can accurately achieve real-time "question-and-answer" dialogues. However, when users engage in multiple rounds of dialogue or conduct business multiple times, or when users resume business processing after a long hiatus, the dialogue system often needs to re-establish the dialogue with the user and repeat past conversations to understand the user's intent or meet their needs. This significantly impacts dialogue and business processing efficiency, hindering the ability to quickly and accurately respond to user dialogues and assist users in processing their business.
[0025] To address the aforementioned technical problems, those skilled in the art have proposed the following typical solutions: 1. Long context window technical solution: This approach accommodates more dialogue history by directly extending the context processing capabilities of the LLM model itself (e.g., expanding the context window from 4K tokens to 128K, 200K, or even longer). The technical approach is to increase the total amount of text that can be processed in a single inference, attempting to use longer historical dialogues directly as the model's input context, thus enabling the model to refer to historical dialogues when responding to the user's current dialogue.
[0026] 2. Retrieval Enhancement Generation (RAG) Technology Solution Based on Vector Database: This approach does not alter the LLM model itself, but instead introduces an external storage and retrieval mechanism. The technical approach is as follows: historical dialogue records are stored entirely or in segments in a vector database; when a new user dialogue arrives, the semantic similarity between the new dialogue and historical dialogue records is calculated, and the most relevant historical dialogue fragments are retrieved from the vector database; these retrieved fragments are then used as additional context and input into the LLM along with the current query, so that the model can refer to historical information to provide an answer.
[0027] 3. Traditional conversation memory technology solutions: This approach is typically implemented at the application layer, where the conversation history is temporarily stored in server memory or cache within the lifecycle of a single session. As long as the session has not timed out or been reset, the agent can access the history from this session as context when responding.
[0028] While the above solutions can improve the dialogue efficiency and accuracy of intelligent dialogue systems to some extent, each still has its own drawbacks: 1. Disadvantages of the long context window scheme: High cost: Processing extremely long contexts significantly increases the consumption of computing resources and inference latency, leading to a sharp rise in service costs.
[0029] Inefficiency: As the context length increases, the model's ability to accurately locate and utilize the most critical information from massive amounts of data decreases, affecting the accuracy and reliability of the response.
[0030] Non-persistent: This scheme can only extend the "memory" of a single conversation. Once the conversation ends, all contextual information is discarded, and long-term memory across sessions cannot be achieved.
[0031] 2. Disadvantages of vector database retrieval schemes: Lack of dynamic state management: This solution excels at handling static knowledge content, but struggles to effectively represent and update dynamically changing dialogue states (such as current business processing progress, to-do list, temporary decisions in the conversation, etc.), resulting in a lack of coherent task orientation in the agent's behavior.
[0032] Inaccurate search results: Semantic similarity-based searches may return incomplete, irrelevant, or redundant information, failing to guarantee strict behavioral consistency in multiple rounds of complex interactions, and easily leading to information confusion or logical conflicts.
[0033] Low degree of information structuring: The stored and retrieved information is raw or simply segmented text, lacking a structured understanding and organization of the information's inherent logic (such as causal relationships, task steps, and user preferences).
[0034] 3. Disadvantages of traditional conversation memory schemes: Short memory period: Its memory is strictly limited to a single session. Once the session is terminated due to timeout, user closing the browser or changing devices, all "memories" will be completely lost, making it impossible to build long-term interactive relationships across multiple sessions.
[0035] Therefore, none of the above solutions can achieve an ideal and efficient intelligent dialogue response.
[0036] To address the aforementioned technical problems, this application provides a dialogue processing method. This method structurally stores the historical dialogue content of users handling business to obtain historical business records. Then, when engaging in dialogue with a user regarding a target business, it retrieves structured historical business record information from the business record database and generates the response content for the current dialogue based on the retrieved historical business record information. This solution provides more refined and persistent storage of historical dialogue records, improving the storage efficiency of useful historical information. When using historical dialogue content to assist the current dialogue, only the structured information extracted from the historical dialogue content needs to be used, rather than directly inputting the original historical dialogue text into the model. This reduces context input overhead, improving dialogue accuracy and efficiency with minimal overhead.
[0037] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0038] The dialogue processing method provided in this application can be applied to an intelligent dialogue system, which can be a subsystem built into a functional system such as an intelligent business processing system or an intelligent customer service system. When a user conducts business processing or makes business inquiries in the aforementioned intelligent business system, intelligent customer service system, or other functional systems, the functional system can invoke the intelligent dialogue system to engage in dialogue with the user, thereby assisting the user in processing business or answering the user's inquiries.
[0039] The aforementioned intelligent dialogue system can be a large language model-based system. This system utilizes the large language model to achieve functions such as user dialogue understanding and dialogue response generation. Furthermore, it can include modules for speech or image recognition and speech or image synthesis, enabling the recognition of multimodal information input by the user and the output of multimodal information to the user. The system also includes a storage unit to house a business record database. This database stores historical business records for different users according to different business categories. For example, a sub-database can be set up for each business, storing the historical data of each user's transactions for that business at the user level.
[0040] See Figure 1 As shown, the dialogue processing method provided in this application includes: S101. Obtain the current dialogue content input by the user.
[0041] The current dialogue content mentioned above includes information related to handling the target business.
[0042] The target business mentioned above can be any business supported by the intelligent dialogue system, such as loan processing, insurance processing, pension processing, etc.
[0043] During the dialogue between the user and the intelligent dialogue system, the user can consult or handle any target business.
[0044] For example, a user can type "I want to apply for a housing loan, how do I do it?" into the intelligent dialogue system. At this time, the intelligent dialogue system can identify from the current dialogue content that the user's target business is "housing loan".
[0045] In addition, intelligent dialogue systems can determine a user's identity based on the user's input identity information or the user's voice information.
[0046] S102. Retrieve historical business records from the business record database that correspond to the user and are related to the target business.
[0047] Specifically, after determining the user's identity and the target business involved in the user's dialogue through step S101, the intelligent dialogue system retrieves historical business records related to the target business and corresponding to the user from the business record database. That is, it retrieves whether the user has handled the target business through the intelligent dialogue system from the business record database and retrieves relevant record information.
[0048] In this embodiment, the business record database stores the user's historical business records in a structured manner, that is, it stores the user's historical business records in two categories: business memory information and background knowledge information.
[0049] The aforementioned business memory information and background knowledge information are extracted from the user's historical dialogues with the intelligent dialogue system.
[0050] The aforementioned business memory information refers to information related to the user's historical records of handling the target business. Specifically, it includes summaries of the user's historical dialogues with the intelligent dialogue system to handle the target business, as well as the latest progress and / or status of the user's handling of the target business as determined by the user's historical dialogues with the intelligent dialogue system, such as the completed steps and pending items of the user's handling of the target business.
[0051] The aforementioned background knowledge information refers to static knowledge information related to the business policies of users and / or target businesses, specifically including user profile information and / or target business policy information.
[0052] The user profile information can include basic personal information such as name, gender, age, and occupation, as well as personalized information such as user hobbies, communication style, and past transactions.
[0053] In some embodiments, it is disclosed that during each historical dialogue between the user and the intelligent dialogue system, for each round of historical dialogue between the user and the intelligent dialogue system, the aforementioned business memory information and background knowledge information are identified, and then the identified business memory information and background knowledge information corresponding to the same user and the same business in the business record database are updated or supplemented using the identified business memory information and background knowledge information.
[0054] For example, the dialogue between the user and the intelligent dialogue system can be input into a large language model, which can then identify business memory information and background knowledge information from it.
[0055] Through the above processing, business memory information can be updated in real time and background knowledge information can be continuously accumulated during the dialogue between the user and the intelligent dialogue system. This allows the business memory information to not only fully remember the summary information of the user's historical dialogue with the intelligent dialogue system, but also to record the latest progress and / or status of the user's business transactions in real time. At the same time, the above processing can continuously accumulate background knowledge information of the user's business transactions, making the recording of user profiles and business policies increasingly more complete.
[0056] In other embodiments, it is disclosed that prior to step S101 above, the user has communicated with the intelligent dialogue system regarding the target business. During this historical communication process, the intelligent dialogue system extracts historical business records corresponding to the user and related to the target business from the historical dialogue content of the user's handling of the target business, and stores the historical business records in the business record database.
[0057] When extracting historical business records that correspond to the user and are related to the target business from the aforementioned historical dialogue content, the process first involves summarizing the historical dialogue content of the user's handling of the target business to obtain a summary of the historical dialogue, determining the progress and / or status of the user's handling of the target business based on the historical dialogue content, and extracting the user's profile information based on the historical dialogue content.
[0058] For example, a dialogue understanding model can be pre-trained to extract dialogue summaries, the user's progress and / or status in handling the target business, and user profile information from the aforementioned historical dialogue content. This dialogue understanding model can employ a large language model.
[0059] The intelligent dialogue system inputs the aforementioned historical dialogue content into the trained dialogue understanding model, and can then extract dialogue summaries, the user's progress and / or status in handling the target business, and user profile information from the historical dialogue content.
[0060] In addition, the intelligent dialogue system can also obtain business policy information of the target business mentioned in the above-mentioned historical dialogue with the user, as background knowledge information for communicating with the user.
[0061] After extracting the above information, the intelligent dialogue system stores the obtained summary of the historical dialogue and the progress and / or status of the user's handling of the target business as business memory information in the historical business record, and stores the obtained user profile information and the business policy information of the target business as background knowledge information in the historical business record.
[0062] After each round of historical dialogue or each historical dialogue between the user and the intelligent dialogue system regarding the target business is completed, the dialogue summary, the user's progress and / or status in handling the target business, and the user's profile information are extracted through the above processing. The extracted information is then used to update the business memory information and background knowledge information in the business record database.
[0063] Through the above processing, the intelligent dialogue system in this embodiment achieves persistence of user history dialogue and continuous storage across sessions. This method breaks through the limitation of session cycle, enabling business status and user information to be preserved for a long time, ensuring the behavioral consistency and goal consistency of the intelligent dialogue system in cross-session interaction.
[0064] Meanwhile, this embodiment divides user history dialogues into business memory information and background knowledge information for structured storage, which makes the retrieval and updating of history dialogue content more efficient and accurate.
[0065] Based on the business record database constructed and maintained through the above processing, when the intelligent dialogue system obtains the current dialogue content input by the user, it retrieves historical business records from the business record database that are related to the target business involved in the current dialogue content and that correspond to the user.
[0066] For example, the intelligent dialogue system retrieves historical business records corresponding to the aforementioned user from a sub-database used to store relevant information about the target business. In other words, it retrieves historical business records extracted from historical dialogues conducted by the aforementioned user through the intelligent dialogue system to handle the target business.
[0067] During the above search process, historical business records can be retrieved through one or more of the following methods: semantic search, keyword search, vector matching, etc.
[0068] S103. If a historical business record corresponding to the user and related to the target business is retrieved, a reply content corresponding to the current dialogue content is generated based on the historical business record and the current dialogue content.
[0069] Specifically, if a historical business record corresponding to the aforementioned user and related to the target business is retrieved from the business record database, the intelligent dialogue system will use this historical business record as reference information, combine it with the current dialogue content input by the user, and generate a response to the user's current dialogue content.
[0070] In some embodiments, when an intelligent dialogue system generates a response to a user's current dialogue content based on historical business records and the current dialogue content, it first generates dialogue prompts using historical business records and the current dialogue content.
[0071] Then, the dialogue prompt is input into the large language model, so that the large language model can generate a response corresponding to the current dialogue content based on the historical business records in the dialogue prompt and the current dialogue content.
[0072] The intelligent dialogue system assembles retrieved historical business records, system preset commands, and current dialogue content input by the user according to a predefined template to obtain dialogue prompts.
[0073] The aforementioned system preset instructions are used to instruct the large language model to generate response content corresponding to the current dialogue content based on historical business records and the current dialogue content.
[0074] In another embodiment, the aforementioned system preset instruction is also used to instruct the large language model to determine whether it can accurately generate a response content corresponding to the current dialogue content based on historical business records and current dialogue content. If it is found that the necessary information for generating the response content is missing, a follow-up question is generated based on the missing information and output. After obtaining the user's answer to the follow-up question, the response content corresponding to the current dialogue content is generated based on historical business records, current dialogue content, and the user's answer to the follow-up question.
[0075] As described above, the dialogue processing method provided in this application extracts business memory information and background knowledge information from the user's historical dialogues related to the target business and stores them as historical business records. When the user inputs current dialogue content related to the target business, the method retrieves historical business records corresponding to the user and related to the target business from the business record database, that is, it retrieves the user's business memory information and background knowledge information related to the target business. Then, based on the retrieved historical business records and the current dialogue content, the method generates a response content corresponding to the current dialogue content.
[0076] The aforementioned method extracts and stores the user's historical dialogues for the target service into business memory information and background knowledge information, achieving structured storage of historical dialogue content. This means it enables persistent and continuous storage of user's historical dialogues across sessions, overcoming the limitations of session cycles and allowing for long-term preservation of business status and user information. Furthermore, this embodiment stores user historical dialogues in a structured manner, dividing them into business memory information and background knowledge information. This makes the retrieval and updating of historical dialogue content more efficient and accurate, avoiding the waste of storing redundant information and preventing redundant information from negatively impacting subsequent dialogue responses.
[0077] Based on the above characteristics, the embodiments of this application use the structured historical business information to respond to dialogues, which can improve the accuracy of responses, improve dialogue efficiency, and improve the efficiency of users in handling business.
[0078] Another embodiment discloses that after generating a response based on historical business records and the current dialogue content, the response content is output. For example, the response content can be output in the form of text or voice.
[0079] Furthermore, in this embodiment, after generating the response content corresponding to the current dialogue content based on historical business records and the current dialogue content according to the above scheme, the business memory information and / or background knowledge information in the historical business records are also updated based on the current dialogue content and / or the response content.
[0080] Specifically, after generating a response based on historical business records and current conversation content, the latest round of conversation with the user is completed. Based on this latest round of conversation, this embodiment continues to update the user's historical business records.
[0081] During the update process, you can refer to the following steps A1-A3 for execution: A1. Extract a dialogue summary and the user's profile information based on the current dialogue content and / or the reply content, and determine the latest progress and / or the latest status of the user's handling of the target business based on the current dialogue content and / or the reply content.
[0082] Specifically, the current dialogue content and / or response content are first summarized to obtain a dialogue summary, and user profile information is extracted from it. Then, based on the current dialogue content and / or response content, the latest progress and / or status of the user's target business is determined.
[0083] For example, a dialogue understanding model can be pre-trained to extract dialogue summaries, the user's progress and / or status in handling the target business, and user profile information from the dialogue. The aforementioned dialogue understanding model can employ a large language model.
[0084] The intelligent dialogue system inputs the current dialogue content and / or response content into the trained dialogue understanding model, and can then extract a dialogue summary, the user's progress and / or status in handling the target business, and the user's profile information from the current dialogue content and / or response content.
[0085] In addition, the intelligent dialogue system can also obtain business policy information of the target business referenced when generating the response content corresponding to the user's current dialogue content, as background knowledge information for communicating with the user.
[0086] A2. Store the extracted dialogue summary in the business memory information in the historical business record, and update the progress and / or status of the user's handling of the target business stored in the business memory information using the latest progress and / or status of the user's handling of the target business.
[0087] Specifically, the dialogue summary and the latest progress and / or status of the user's target service obtained in step A1 can be directly used to update the service memory information in the historical service records, thereby achieving an overlay update of the service memory information. For example, the dialogue summary extracted from the current dialogue content and / or the corresponding reply content can be stored in the service memory information as a historical dialogue summary of the user's target service; the latest progress and / or status of the user's target service can be used to update the progress and / or status of the user's target service stored in the service memory information, ensuring that the progress and / or status of the user's target service stored in the service memory information remains up-to-date.
[0088] A3. If the extracted user profile information and / or the target business policy information are updated relative to the user profile information and / or business policy information stored in the background knowledge information of the historical business records, the user profile information and / or business policy information stored in the background knowledge information of the historical business records shall be updated using the extracted user profile information and / or the updated business policy information of the target business.
[0089] Specifically, background knowledge information in historical business records is updated cumulatively. That is, for the user profile information obtained in step A1 and / or the business policy information of the target business referenced in the generated response content corresponding to the user's current dialogue content, if the user profile information and / or the business policy information of the target business obtained in step A1 are updated compared to the user profile information and / or the business policy information stored in the background knowledge information of historical business records—for example, if step A1 obtains new user profile information, or the generated response content references new business policy information of the target business—then the user profile information and / or the business policy information stored in the background knowledge information of historical business records are updated using the newly obtained user profile information and / or the new business policy information of the target business. In other words, the newly obtained user profile information and / or the new business policy information of the target business are added to the background knowledge information of historical business records.
[0090] Furthermore, in this embodiment, if the business policy information of the target business stored in the background knowledge information of the aforementioned historical business records is updated, even if there is no dialogue between the user and the intelligent dialogue system, the business policy information of the target business stored in the background knowledge information can be directly updated to the latest business policy information.
[0091] The solution in this embodiment can continuously update and supplement the user's historical business records each time the user has a dialogue with the intelligent dialogue system on the target business, thereby realizing long-term and persistent recording of dialogue content across sessions. This provides continuously updated historical dialogue information to support future user dialogue responses, thereby continuously improving dialogue quality and efficiency.
[0092] The following example illustrates the processing procedure of the dialogue processing method provided in this application.
[0093] Suppose a user applies for "insurance enrollment for flexible employment personnel" through an intelligent dialogue system.
[0094] In the first conversation, Mr. Wang asked, "I would like to know how to enroll in social insurance for flexible employment personnel." At this point, the intelligent dialogue system understands the user's inquiry and checks the business policy for "flexible employment personnel insurance enrollment" to determine that the insured person needs to provide their ID card and bank card. The intelligent dialogue system then replies to Mr. Wang: "Hello, to apply for flexible employment personnel insurance enrollment, the insured person needs to prepare their ID card and bank card." Based on the initial dialogue, the intelligent dialogue system analyzes the content and determines the dialogue summary as "User Goal: Apply for flexible employment insurance." The current status is "Policy consultation has been conducted, and the user has been informed that an ID card and bank card are required." The user profile for Mr. Wang is "Business Intent: Flexible employment insurance." The intelligent dialogue system then stores the dialogue summary and current status as business memory information for Mr. Wang's historical business records, and stores the user profile as background knowledge information for his historical business records.
[0095] A few days later, Mr. Wang came again to speak with the intelligent dialogue system in order to continue processing the "flexible employment personnel insurance" business.
[0096] In the second conversation, Mr. Wang typed: "My materials are ready." At this point, after the intelligent dialogue system identifies the user Mr. Wang, it queries the business record database for historical business records that correspond to Mr. Wang and are related to the business involved in "My materials are ready."
[0097] The search reveals Mr. Wang's business history in the business record database, including the dialogue summary: "User goal: Apply for flexible employment insurance." Current status: "Policy consultation has been conducted, and he has been informed that he needs to prepare his ID card and bank card." Mr. Wang's profile: "Business intention: Flexible employment insurance." At this point, the intelligent dialogue system, combining Mr. Wang's business history with his input of "My materials are ready," generates a response: "Hello sir, it's a pleasure to continue serving you. Are you preparing to register for flexible employment social insurance now?" Clearly, the intelligent dialogue system's response directly connects to past business transactions, eliminating the need to repeatedly ask the user for their intentions.
[0098] During the subsequent dialogue, the intelligent dialogue system continued to analyze the dialogue content between user Mr. Wang and the intelligent dialogue system in real time, and updated the business memory information and background knowledge information of the corresponding user Mr. Wang stored in the business record database based on the information obtained during the communication process.
[0099] For example, when Mr. Wang completes the "flexible employment personnel insurance" business through the intelligent dialogue system, the "current status" in the business memory information is updated after each step, such as "Completed: Identity information registration. Pending: Material upload and confirmation". During the conversation with user Mr. Wang, the intelligent dialogue system can automatically adjust the reply to more concise command language based on the profile information of user Mr. Wang recorded in the background knowledge information, which shows that he "prefers concise communication", thereby meeting the user's communication preferences.
[0100] Corresponding to the dialogue processing method described above, this application also provides a dialogue processing apparatus, see below. Figure 2 As shown, the device includes: The dialogue acquisition unit 100 is used to acquire the current dialogue content input by the user, which includes content related to handling the target business. The data retrieval unit 110 is used to retrieve historical business records corresponding to the user and related to the target business from the business record database; the historical business records include business memory information and background knowledge information, the business memory information includes the user's historical records of handling the target business, and the background knowledge information includes the user's profile information and / or the business policy information of the target business; The dialogue response unit 120 is used to generate response content corresponding to the current dialogue content based on the historical business record corresponding to the user and related to the target business when a historical business record is retrieved.
[0101] In some embodiments, the user's history of handling the target service includes a summary of the user's historical conversations regarding the target service, as well as the progress and / or status of the user's handling of the target service; The user's profile information is extracted based on the user's historical conversations when handling the target service.
[0102] In some embodiments, the dialogue response unit 120 generates response content corresponding to the current dialogue content based on the historical business records and the current dialogue content, including: Using the historical business records and the current dialogue content, generate dialogue prompts; The dialogue prompts are input into a large language model, which then generates a response based on the historical business records and the current dialogue content.
[0103] In some embodiments, the apparatus further includes: An information extraction unit is used to extract historical business records that correspond to the user and are related to the target business from the user's historical dialogue content when the user handles the target business. The historical business records are stored in the business record database.
[0104] In some embodiments, the information extraction unit extracts historical business records corresponding to the user and related to the target business from the user's historical dialogue content when handling the target business, including: The process involves summarizing the historical dialogue content of the user's handling of the target service to obtain a summary of the historical dialogue, determining the progress and / or status of the user's handling of the target service based on the historical dialogue content, and extracting the user's profile information based on the historical dialogue content. The summary of the historical dialogue and the progress and / or status of the user's handling of the target business are stored as business memory information in the historical business record. The user's profile information and the business policy information of the target business are stored as background knowledge information in the historical business record.
[0105] In some embodiments, the apparatus further includes: The information update unit is used to update the business memory information and / or the background knowledge information in the historical business records based on the current dialogue content and / or the reply content.
[0106] In some embodiments, the information updating unit updates the business memory information and / or the background knowledge information in the historical business records based on the current dialogue content and / or the response content, including: Based on the current dialogue content and / or the reply content, extract the dialogue summary and the user's profile information, and based on the current dialogue content and / or the reply content, determine the latest progress and / or the latest status of the user's handling of the target business; The extracted dialogue summary is stored in the business memory information in the historical business record, and the progress and / or status of the user's handling of the target business stored in the business memory information is updated using the latest progress and / or status of the user's handling of the target business. If the extracted user profile information and / or the target business policy information are updated relative to the user profile information and / or business policy information stored in the background knowledge information of the historical business records, the extracted user profile information and / or the updated business policy information of the target business shall be used to update the user profile information and / or business policy information stored in the background knowledge information of the historical business records.
[0107] The dialogue processing apparatus provided in this embodiment belongs to the same concept as the dialogue processing method provided in the above embodiments of this application. It can execute the dialogue processing method provided in any of the above embodiments of this application and has the corresponding functional modules and beneficial effects of the method. Technical details not described in detail in this embodiment can be found in the specific processing content of the dialogue processing method provided in the above embodiments of this application, and will not be repeated here.
[0108] The functions implemented by each of the above units can be implemented by the same or different processors, and this application embodiment does not limit this.
[0109] It should be understood that the units in the above device can be implemented by a processor calling software. For example, the device includes a processor connected to a memory containing instructions. The processor calls the instructions stored in the memory to implement any of the above methods or to implement the functions of each unit in the device. The processor can be a general-purpose processor, such as a CPU or microprocessor, and the memory can be internal or external to the device. Alternatively, the units in the device can be implemented as hardware circuits. By designing the hardware circuits, some or all of the unit functions can be implemented. The hardware circuits can be understood as one or more processors. For example, in one implementation, the hardware circuit is an ASIC, and the functions of some or all of the above units are implemented by designing the logical relationships between the components within the circuit. In another implementation, the hardware circuit can be implemented using a PLD, such as an FPGA, which can include a large number of logic gates. The connection relationships between the logic gates are configured through configuration files to implement the functions of some or all of the above units. All units in the above device can be implemented entirely by a processor calling software, entirely by hardware circuits, or partially by a processor calling software with the remaining parts implemented by hardware circuits.
[0110] In this application embodiment, a processor is a circuit with signal processing capabilities. In one implementation, the processor can be a circuit with instruction reading and execution capabilities, such as a CPU, microprocessor, GPU, or DSP. In another implementation, the processor can implement certain functions through the logical relationships of hardware circuits. These logical relationships are fixed or reconfigurable. For example, the processor may be a hardware circuit implemented as an ASIC or PLD, such as an FPGA. In a reconfigurable hardware circuit, the process of the processor loading a configuration document and configuring the hardware circuit can be understood as the processor loading instructions to implement the functions of some or all of the above units. Furthermore, it can also be a hardware circuit designed for artificial intelligence, which can be understood as an ASIC, such as an NPU, TPU, or DPU.
[0111] As can be seen, each unit in the above device can be one or more processors (or processing circuits) configured to implement the above methods, such as: CPU, GPU, NPU, TPU, DPU, microprocessor, DSP, ASIC, FPGA, or a combination of at least two of these processor forms.
[0112] Furthermore, the units in the above devices can be integrated in whole or in part, or they can be implemented independently. In one implementation, these units are integrated together and implemented in the form of a System-on-Chip (SoC). The SoC may include at least one processor for implementing any of the above methods or implementing the functions of the units in the device. The at least one processor may be of different types, such as CPU and FPGA, CPU and artificial intelligence processor, CPU and GPU, etc.
[0113] Another embodiment of this application also provides an electronic device, see [link to relevant documentation] Figure 3 As shown, the device includes: Memory 200 and processor 210; The memory 200 is connected to the processor 210 and is used to store programs; The processor 210 is configured to implement the dialogue processing method disclosed in any of the above embodiments by running the program stored in the memory 200.
[0114] Specifically, the aforementioned electronic device may also include: a bus, a communication interface 220, an input device 230, and an output device 240.
[0115] The processor 210, memory 200, communication interface 220, input device 230, and output device 240 are interconnected via a bus. Among them: A bus can include a pathway for transmitting information between various components of a computer system.
[0116] Processor 210 can be a general-purpose processor, such as a general-purpose central processing unit (CPU), a microprocessor, etc., or an application-specific integrated circuit (ASIC), or one or more integrated circuits used to control the execution of the program of the present invention. It can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), an off-the-shelf programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.
[0117] Processor 210 may include a main processor, as well as a baseband chip, modem, etc.
[0118] The memory 200 stores a program that executes the technical solution of this invention, and may also store an operating system and other key business functions. Specifically, the program may include program code, which includes computer operation instructions. More specifically, the memory 200 may include read-only memory (ROM), other types of static storage devices capable of storing static information and instructions, random access memory (RAM), other types of dynamic storage devices capable of storing information and instructions, disk storage, flash memory, etc.
[0119] Input device 230 may include a device for receiving user input data and information, such as a keyboard, mouse, camera, scanner, light pen, voice input device, touch screen, pedometer, or gravity sensor.
[0120] Output device 240 may include devices that allow information to be output to a user, such as a display screen, printer, speaker, etc.
[0121] The communication interface 220 may include a device that uses any transceiver to communicate with other devices or communication networks, such as Ethernet, Radio Access Network (RAN), Wireless Local Area Network (WLAN), etc.
[0122] The processor 210 executes the program stored in the memory 200 and calls other devices, which can be used to implement various steps of any of the dialogue processing methods provided in the above embodiments of this application.
[0123] In addition to the methods and devices described above, embodiments of this application may also be computer program products, which include computer program instructions that, when executed by a processor, cause the processor to perform the steps of the dialogue processing method described in any of the embodiments described above.
[0124] The computer program product can be written in any combination of one or more programming languages to perform the operations of the embodiments of this application. The programming languages include object-oriented programming languages such as Java and C++, as well as conventional procedural programming languages such as C or similar languages. The program code can be executed entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.
[0125] Furthermore, embodiments of this application may also be storage media storing a computer program thereon, which, when run by a processor, causes the processor to execute the steps of the dialogue processing method described in any of the above embodiments of this specification.
[0126] For the foregoing method embodiments, in order to simplify the description, they are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, because according to this application, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to this application.
[0127] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For apparatus embodiments, since they are basically similar to method embodiments, the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.
[0128] The steps in the methods of the various embodiments of this application can be adjusted, merged, or deleted in order according to actual needs, and the technical features described in each embodiment can be replaced or combined.
[0129] The modules and sub-modules in the various embodiments of the present application's devices and terminals can be merged, divided, and deleted according to actual needs.
[0130] It should be understood that the disclosed terminals, devices, and methods can be implemented in other ways, given the several embodiments provided in this application. For example, the terminal embodiments described above are merely illustrative. For instance, the division of modules or sub-modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple sub-modules or modules may be combined or integrated into another module, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or modules, and may be electrical, mechanical, or other forms.
[0131] The modules or submodules described as separate components may or may not be physically separate. The components that constitute a module or submodule may or may not be physical modules or submodules; that is, they may be located in one place or distributed across multiple network modules or submodules. Some or all of the modules or submodules can be selected to achieve the purpose of this embodiment's solution, depending on actual needs.
[0132] Furthermore, the functional modules or sub-modules in the various embodiments of this application can be integrated into one processing module, or each module or sub-module can exist physically separately, or two or more modules or sub-modules can be integrated into one module. The integrated modules or sub-modules described above can be implemented in hardware or in the form of software functional modules or sub-modules.
[0133] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.
[0134] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software unit executed by a processor, or a combination of both. The software unit can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.
[0135] Finally, it should be noted that in this document, relational terms such as "current" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.
[0136] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A dialogue processing method, characterized in that, include: Obtain the current dialogue content input by the user, wherein the current dialogue content includes content related to handling the target business; Retrieve historical business records from the business record database that correspond to the user and are related to the target business; the historical business records include business memory information and background knowledge information, the business memory information includes the user's historical records of handling the target business, and the background knowledge information includes the user's profile information and / or the business policy information of the target business; If historical business records corresponding to the user and related to the target business are retrieved, a response is generated based on the historical business records and the current conversation content.
2. The method according to claim 1, characterized in that, The user's historical records of handling the target service include a summary of the user's historical dialogues regarding the target service, as well as the user's progress and / or status in handling the target service. The user's profile information is extracted based on the user's historical conversations when handling the target service.
3. The method according to claim 1, characterized in that, The step of generating response content corresponding to the current conversation content based on the historical business records and the current conversation content includes: Using the historical business records and the current dialogue content, generate dialogue prompts; The dialogue prompts are input into a large language model, which then generates a response based on the historical business records and the current dialogue content.
4. The method according to any one of claims 1 to 3, characterized in that, The method further includes: Extract historical business records that correspond to the user and are related to the target business from the user's historical dialogue content when the user handles the target business; The historical business records are stored in the business record database.
5. The method according to claim 4, characterized in that, The historical transaction records corresponding to the user and related to the target transaction are extracted from the user's historical dialogue content when the user handles the target transaction, including: The process involves summarizing the historical dialogue content of the user's handling of the target service to obtain a summary of the historical dialogue, determining the progress and / or status of the user's handling of the target service based on the historical dialogue content, and extracting the user's profile information based on the historical dialogue content. The summary of the historical dialogue and the progress and / or status of the user's handling of the target business are stored as business memory information in the historical business record. The user's profile information and the business policy information of the target business are stored as background knowledge information in the historical business record.
6. The method according to any one of claims 1 to 3, characterized in that, The method further includes: Based on the current dialogue content and / or the response content, update the business memory information and / or the background knowledge information in the historical business records.
7. The method according to claim 6, characterized in that, The step of updating the business memory information and / or background knowledge information in the historical business records based on the current dialogue content and / or the response content includes: Based on the current dialogue content and / or the reply content, extract the dialogue summary and the user's profile information, and based on the current dialogue content and / or the reply content, determine the latest progress and / or the latest status of the user's handling of the target business; The extracted dialogue summary is stored in the business memory information in the historical business record, and the progress and / or status of the user's handling of the target business stored in the business memory information is updated using the latest progress and / or status of the user's handling of the target business. If the extracted user profile information and / or the target business policy information are updated relative to the user profile information and / or business policy information stored in the background knowledge information of the historical business records, the extracted user profile information and / or the updated business policy information of the target business shall be used to update the user profile information and / or business policy information stored in the background knowledge information of the historical business records.
8. A dialogue processing device, characterized in that, include: The dialogue acquisition unit is used to acquire the current dialogue content input by the user, which includes content related to handling the target business. The data retrieval unit is used to retrieve historical business records corresponding to the user and related to the target business from the business record database; the historical business records include business memory information and background knowledge information, the business memory information includes the user's historical records of handling the target business, and the background knowledge information includes the user's profile information and / or the business policy information of the target business; The dialogue response unit is used to generate response content corresponding to the current dialogue content based on the historical business records that correspond to the user and are related to the target business, when historical business records are retrieved.
9. An electronic device, characterized in that, Including memory and processor; The memory is connected to the processor and is used to store programs; The processor is used to implement the dialogue processing method as described in any one of claims 1 to 7 by running a program in the memory.
10. A computer program product, characterized in that, It includes computer program instructions that, when executed by a processor, cause the processor to perform the dialogue processing method as described in any one of claims 1 to 7.