CN-122019764-A - Information processing method and system in AI dialogue system and electronic equipment

CN122019764ACN 122019764 ACN122019764 ACN 122019764ACN-122019764-A

Abstract

The embodiment of the application discloses an information processing method, an information processing system and electronic equipment in an AI dialogue system, which comprise the steps of acquiring user archive information through a memory system database associated with the dialogue system after receiving dialogue content input by a user, adding the user archive information into prompt information and providing the prompt information to an AI model so that the AI model generates personalized reply content by referring to the user archive information, providing the personalized reply content generated by the AI model to a client so as to output the personalized reply content through a dialogue interface of the client and displaying a reference information card, wherein the reference information card is used for displaying user archive information referred by the AI model in the process of generating the reply content, and updating the user archive information in the database if dialogue content used for updating the user archive information and input by the user is received. By the embodiment of the application, the conversation effect can be improved, and the trust of the user to the AI system can be enhanced.

Inventors

DONG HUI
CHENG SIQI
CHEN LU
ZHOU CHENGYUE

Assignees

杭州阿里巴巴海外互联网产业有限公司

Dates

Publication Date: 20260512
Application Date: 20251217

Claims (18)

1. An information processing method in an AI dialogue system, comprising: After receiving dialogue content input by a user, acquiring user file information through a memory system database associated with the dialogue system, wherein the user file information is extracted from the dialogue content by an artificial intelligence AI model and dynamically updated into the database in the process of carrying out historical dialogue with the user by the AI model; Adding the user profile information into prompt information and providing the prompt information to an AI model so that the AI model generates personalized reply content by referring to the user profile information; And providing the personalized reply content generated by the AI model to a client so as to be output through a dialogue interface of the client, and displaying a reference information card, wherein the reference information card is used for displaying user archive information referred by the AI model in the process of generating the reply content, and if dialogue content input by a user for updating the user archive information is received, updating the user archive information in the database.
2. The method of claim 1, wherein the step of determining the position of the substrate comprises, In the process of carrying out dialogue with a user through the AI model, historical dialogue content is also saved in a database of the memory system, the memory system comprises a short-term memory cache and a long-term memory library, the short-term memory cache is used for caching dialogue content of the latest preset number of rounds, and the long-term memory library is used for carrying out persistence storage on dialogue content before the preset number of rounds; the method further comprises the steps of: After receiving dialogue content input by a user, carrying out intention recognition according to the dialogue content; If the identified user intention is related to the historical dialog content, the related historical dialog content is retrieved from the memory system and input to the AI model through the prompt message so that the AI model can generate reply content by referring to the related historical dialog content.
3. The method as recited in claim 2, further comprising: the reference information card also comprises historical dialogue content information which is referred by the AI model when generating reply content.
4. The method of claim 2, wherein the step of determining the position of the substrate comprises, When the history dialogue content is stored in the long-term memory bank, the original text of the history dialogue content is stored in an original text data table in the long-term memory bank, so that the AI model can generate reply content by referring to the original text of the related history dialogue content.
5. The method of claim 4, wherein the step of determining the position of the first electrode is performed, When the historical dialogue content is stored in the long-term memory bank, generating summary content for topics which can be repeatedly used and storing the summary content into a summary data table in the long-term memory bank, wherein the summary content is generated by summarizing or summarizing a plurality of dialogue content texts related to the same topic; The retrieving related historical dialog content from the memory system includes: And retrieving summary content of related historical dialogue content from the summary data table and inputting the summary content into the AI model.
6. The method as recited in claim 5, further comprising: After retrieving the summary content, judging whether the historical dialogue content text still needs to be referred to, and if so, retrieving the relevant historical dialogue content text from the text data table for reference by the AI model when generating the reply content.
7. The method of claim 6, wherein the step of providing the first layer comprises, And acquiring user file information from the long-term memory library through an arrangement module, carrying out intention recognition on dialogue content input by a user, determining decision logic according to the intention recognition result and carrying out flow control, wherein the decision logic is used for determining a retrieval mode of a multi-class data table in the memory system.
8. The method of claim 7, wherein the step of determining the position of the probe is performed, The decision logic comprises the steps of judging whether a short-term cache or a long-term memory library is required to be searched to acquire historical dialogue contents, further determining whether a summary or an original text of related historical dialogue contents is required to be acquired if the long-term memory library is required to be searched, and performing flow control according to the decision logic, wherein in the flow control process, a primary text data table and/or a summary data table in the short-term cache or the long-term memory library is searched to acquire dialogue contents in the last preset number of rounds, and after the summary and/or the original text of the dialogue contents before the preset number of rounds of the related historical dialogue contents are acquired, generating prompt information by combining user file information to be input to the AI model.
9. The method as recited in claim 7, further comprising: And determining whether additional information outside the dialogue system is needed or not according to the identified user intention by the arrangement module, and if so, acquiring the additional information by calling a related tool and providing the additional information for the AI model for reference.
10. An information processing method in an AI dialogue system, comprising: Receiving dialogue content input by a user, submitting the dialogue content to a server so that the server obtains user file information through a memory system database associated with the dialogue system and adds the user file information to prompt information, and providing the user file information to an AI model, wherein the user file information is extracted from the dialogue content and dynamically updated to the database by the AI model in the process of carrying out historical dialogue with the user through the AI model, and the AI model is used for generating personalized reply content by referring to the user file information; Outputting the personalized reply content through a dialogue interface of the client, and displaying a reference information card, wherein the reference information card is used for displaying user file information referred by the AI model in the process of generating the reply content; And if receiving dialogue content which is input by a user and is used for updating the user archive information, submitting the dialogue content to a server so as to update the user archive information in the database.
11. An information processing method in an AI dialogue system, comprising: In the process of carrying out historical dialogue with a user through an AI model, adding dialogue contents with the latest preset number of rounds into a short-term memory cache of the AI model, storing the dialogue contents before the preset number of rounds into a long-term memory bank outside the AI model in a lasting manner, extracting user file information from the dialogue contents, storing the dialogue contents into the long-term memory bank in a lasting manner, and dynamically updating the dialogue contents; After receiving the current dialogue content input by a user, acquiring the user file information from the long-term memory bank, and after carrying out intention recognition according to the current dialogue content, if the recognized user intention is related to the history dialogue content, retrieving the related history dialogue content from the short-term memory cache and/or the long-term memory bank; and generating prompt information according to the current dialogue content, the user file information and the related historical dialogue content, and calling an AI model based on the prompt information so that the AI model generates personalized reply content.
12. The method as recited in claim 11, further comprising: And providing the personalized reply content generated by the AI model to a client so as to be output through a dialogue interface of the client, and displaying a reference information card, wherein the reference information card is used for displaying user archive information referred by the AI model in the process of generating the reply content, and if dialogue content input by a user for updating the user archive information is received, updating the user archive information in the database.
13. An information processing system in an AI dialog system, comprising: The memory module comprises a short-term memory cache and a long-term memory library, wherein the short-term memory cache is used for caching conversation contents of the latest preset number of rounds, the long-term memory library comprises an original text data table, a summary data table and a user file data table, the original text data table is used for carrying out lasting storage on conversation content original text before the preset number of rounds, the summary data table is used for carrying out lasting storage on summary contents generated after summarizing or related to topics which can be repeatedly used, and the user file data table is used for carrying out lasting storage on user file information, and the user file information is extracted from the conversation contents and dynamically updated into the user file data table by an AI model in the process of carrying out history conversation with a user through an artificial intelligent AI model; And the arrangement module acquires user file information from a user file data table of the long-term memory bank after receiving dialogue content input by a user, carries out intention recognition on the dialogue content input by the user, and determines decision logic according to the intention recognition result, wherein the decision logic comprises determining whether a short-term cache or a long-term memory bank needs to be searched to acquire history dialogue content, if the long-term memory bank needs to be searched, further determining whether a summary or an original text of related history dialogue content needs to be acquired, so as to search the original text data table and/or the summary data table in the short-term cache, the long-term memory bank according to the decision logic, and generate prompt information by combining the user file information after acquiring dialogue content of the related history dialogue content in a last preset number of rounds, the summary and/or the original text of the dialogue content before the preset number of rounds, and then input the prompt information into an AI model, so that the AI model generates personalized reply content by referring to the user file information and the related history dialogue content.
14. The system of claim 13, wherein the system further comprises a controller configured to control the controller, The arrangement layer is further used for outputting the personalized reply content through a dialogue interface of the client, displaying a reference information card, wherein the reference information card is used for displaying user file information referred by the AI model in the process of generating the reply content, and if dialogue content input by a user for updating the user file information is received, updating the user file information in the user file data table.
15. The system of claim 13, wherein the system further comprises a controller configured to control the controller, The arrangement module is further used for determining whether additional information outside the dialogue system is needed according to the identified user intention, and if so, acquiring the additional information in a mode of calling a related tool and providing the additional information for the AI model for reference.
16. A computer readable storage medium, on which a computer program is stored, characterized in that the program, when being executed by a processor, implements the steps of the method of any of claims 1 to 12.
17. An electronic device, comprising: One or more processors, and A memory associated with the one or more processors for storing program instructions that, when read for execution by the one or more processors, perform the steps of the method of any of claims 1 to 12.
18. A computer program product comprising computer program/computer-executable instructions which, when executed by a processor in an electronic device, implement the steps of the method of any one of claims 1 to 12.

Description

Information processing method and system in AI dialogue system and electronic equipment Technical Field The present application relates to the field of information processing technologies, and in particular, to an information processing method and system in an AI dialogue system, and an electronic device. Background In recent years, AI (ARTIFICIAL INTELLIGENCE ) models have made significant progress in dialog systems, particularly "Wen Shengwen" class language models, as powerful "text-to-speech" models, AI models can play a specific role (e.g., customer service, technical support, etc.) by pre-designed prompt words, and generate reasonable, consistent reply content given a piece of dialog content (including the currently received dialog content, and the dialog history as context). However, the AI model still presents challenges in the process of application in the dialog system described above. For example, the dialog model, at a glance for all users, lacks long-term memory of the identity, preferences, and history of a particular user, and thus does not provide truly personalized services. Each user has no distinction in the model eye, the answer lacks pertinence, and the customized experience desired by the user is difficult to meet. One technical attempt to address the above problems is to fix the user representation data, i.e., create a file (e.g., fill out a questionnaire or import material, etc.) for the user in advance for the dialog model to read in the process of generating the reply content. The method realizes personalization to a certain extent, but the user portrait data is usually manually maintained and is difficult to update in real time, so that the personalization is insufficient. Disclosure of Invention The application provides an information processing method, an information processing system and electronic equipment in an AI dialogue system, which can enhance the trust of a user to the AI system while improving dialogue effect. The application provides the following scheme: an information processing method in an AI conversation system, comprising: After receiving dialogue content input by a user, acquiring user file information through a memory system database associated with the dialogue system, wherein the user file information is extracted from the dialogue content by an artificial intelligence AI model and dynamically updated into the database in the process of carrying out historical dialogue with the user by the AI model; Adding the user profile information into prompt information and providing the prompt information to an AI model so that the AI model generates personalized reply content by referring to the user profile information; And providing the personalized reply content generated by the AI model to a client so as to be output through a dialogue interface of the client, and displaying a reference information card, wherein the reference information card is used for displaying user archive information referred by the AI model in the process of generating the reply content, and if dialogue content input by a user for updating the user archive information is received, updating the user archive information in the database. In the process of carrying out dialogue with a user through the AI model, historical dialogue content is also saved in a database of the memory system, the memory system comprises a short-term memory cache and a long-term memory library, the short-term memory cache is used for caching dialogue content of the latest preset number of rounds, and the long-term memory library is used for carrying out persistence storage on dialogue content before the preset number of rounds; the method further comprises the steps of: After receiving dialogue content input by a user, carrying out intention recognition according to the dialogue content; If the identified user intention is related to the historical dialog content, the related historical dialog content is retrieved from the memory system and input to the AI model through the prompt message so that the AI model can generate reply content by referring to the related historical dialog content. Wherein, still include: the reference information card also comprises historical dialogue content information which is referred by the AI model when generating reply content. When the history dialogue content is stored in the long-term memory bank, the original text of the history dialogue content is stored in an original text data table in the long-term memory bank, so that the AI model can generate reply content by referring to the original text of the related history dialogue content. When the historical dialogue content is stored in the long-term memory bank, generating summary content for topics possibly used repeatedly, and storing the summary content in a summary data table in the long-term memory bank, wherein the summary content is generated by summarizing or summarizing a plurality of dialogue content texts related to th